Tuesday, December 18, 2012

Monday, January 9, 2012

ZINC (Horsehead) Investment Thesis

Horsehead Holding Corp (ZINC) is a company that recently came to my attention. ZINC is a metal waste recycler with a turbulent recent past, and it appears to be one of the few zinc pure plays for an investor looking to bet on zinc prices.

Horsehead's 3 primary lines of business are:
  1. Collecting EAF dust - it carts off EAF dust (a zinc-rich waste byproduct) from steel mini mlls for a fee. It claims to be the largest EAF dust recycler in the United States. Mini mills today can either sell their EAF dust waste to recyclers like Horsehead or pay landfills to bury it under the ground.

  2. Recycling zinc bearing waste (primarily EAF dust) and converting it into commodity Zinc oxide and PW zinc metal, which it then sells at prices that track the market price of zinc. It claims to be the largest zinc-bearing waste recycler in the United States.

  3. Recycling nickel bearing waste (including NiCd) batteries and converting it into Nickel based products for sale in the market. Nickel is a commodity and the prices Horsehead commands roughly track nickel prices.

The Zinc recycling business accounts for 75% of its revenue and appears to be management's primary focus. In its zinc recycling business, the company buys EAF dust (a waste byproduct of steel manufacturing in mills that use Electric Arc Furnaces aka "mini-mills") and zinc secondaries (other forms of waste that has zinc), and refines them into zinc which it then sells. The bulk of its feedstock is EAF dust, which is collected from Nucor and other "mini-mill" steel manufacturers, and processed as follows:


Conversion Stage 1: The company's Palmerton, Rockwood, and Calumet facilties primarily use Waelz Kiln technology (the technology has been known since the late 1800s, and has been in commercial use since the 1970s) preprocessing EAF dust to produce CZO (55-65% zinc content).

Conversion Stage 2: The CZO is then shipped to the Palmerton facility which uses natural gas to heat it to burn off impurities, leaving behind zinc calcine (65-70% zinc content). The zinc calcine granules are then shipped to the Monaca plant as feedstock and/or sold to other facilities around the world.

Conversion Stage 3: The Monaca plant then refines this feedstock into the finished products: zinc oxide and zinc metal. The Monaca plant can also take in other types of feedstock: (a) CZO (they have installed washing facilities to allow it to receive CZO directly and wash off chlorine impurities), or (b) zinc secondaries (other types of zinc containing waste). The plant also uses metallurgical coke, natural gas and electricity in its operations, and its electricity is solely provided by an onsite 110 megawatt coal-fired power plant that uses Power River Basin coal. The coal contracts appear to be renewed every 1-2 years. The Monaca electrothermic zinc smelter and refinery has a 175,000 ton per year smelting capacity (zinc containing equivalent?), and is able to produce:
- PW metal 88,000 tons per year
- Zinc Oxide 90,000 tons per year
- SSHG Metal 15,000 tons per year
- Zinc Dust 5,900 tons per year

The Monaca plant is scheduled to be replaced by a new refinery in Rutherford County, North Carolina in Q3 2013. The new plant is intended to lower the cost of production by approximately 35%, and to comply with new environmental regulations. (As opposed to its Monaca plant, which will require capital investments to become compliant).


The economics of the business
and its ecological niche in the business ecosystem

While it is tempting to say that Horsehead exists in a mutually beneficial symbiotry with steel mini-mills, this would not be entirely accurate. Why? Because the benefical relationship is one way, and not a "closed loop".

The mini-mills benefit by having lower EAF waste disposal costs, because they can pay Horsehead less than what they pay landfills to take their EAF dust. (in terms of total costs - for example that impute the potentially higher risk of breaching environmental regulations when using landfills). If competition comes in the future, it is entirely possible that Horsehead has to pay the mini-mills for their EAF dust.

However, the minimills do not depend on the zinc metal that Horsehead produces. While they do require zinc for processes such as galvanizing steel, they can and will source for zinc metal from the open market. In selling its zinc metal, Horsehead competes in an open commodity market, and is a price-taker.

The economics of the business are that it is a price taker in a commodity market, and to maintain its position, it needs to maintain a low or have the lowest cost of production. Its profit potential is largely dependent on exogenous factors, namely it is the spread between the cost of EAF dust and the market price of zinc, minus its commodity intensive cost of production. Changes in relative commodity prices could destroy its business model/ecological niche. For example, if the cost of zinc falls and the cost of energy used in recycling rises relative to the cost of energy used in mining, the business of recycling EAF dust could become untenable.




Competitive forces and competitive position

Horsehead's exists in a symbiotry of commensalism with mini-mills in the United States. It's competitive position in the ecosystem, and the existence of its ecological niche, is dependent on: (1) it being able to secure EAF dust from mini-mills by making it cheaper for the mills to send it to them as opposed to the landfills/other recyclers, (2) it being able to produce zinc at the low end of the zinc cost curve compared to mined zinc and other recycled zinc. In the latter, the factors of production in recycling and mining can be a key consideration - for example, if recycling technologies require far less energy than mining, then recycling can have a strong cost advantage in times of high energy prices.

Apart from these horizontal competitors, Horsehead also has to secure its place in the ecosystem from the threat of vertical integration. In other words, why don't the mills process the EAF dust themselves? A mini-mill looking to reduce its cost of operations may choose to run its own EAF dust recycling facility to make some money to offset steel production costs. There are a couple of reasons why they may find it  uneconomical to do so. For example:
  • Horsehead may have a vastly superior proprietary recycling technology protected by patents or trade secrets, which competitors are unable to work around. This appears to be unlikely, since Horsehead has not said that they have a technological ace up their sleeve.

  • There may be economies of scale in recycling, such that a central recycler can achieve much lower costs of production than an individual mini-mill. This is probable, given that nature of industrial processes and the fragmented nature of steel production (i.e. no dominant mini-mills with >40-50% market share that can achieve greater zinc recycling cost efficiency than any 3rd party zinc recycler like Horsehead)

  • The ability to quickly start and stop production (a key advantage that mini-mills have over integrated blast-furnace steel mills) is incompatible with recycling processes. For example, recycling processes may require continuous operation or face costly shut-down / start-up procedures.
Nonethless, a mini-mill operator looking to increase its share of the value chain could decide to get into the EAF dust recycling, or use the threat of this to squeeze the profits out of Horsehead by charging Horsehead for its EAF dust.


Sustainability of the competitive position and the ecological niche (Parameters of an investment thesis)

Given the economics of the business, the key to Horsehead sustaining its earnings stream are likely to be:
  1. The ability to preserve the ecological niche of recycling EAF dust into zinc. This requires that:

    (a) That steel manufacturing using EAF continues to be in vogue; only Electric Arc Furnace mini mills produce EAF dust, which is Horsehead's primary raw material / feedstock.

    (b) The "best" zinc recycling technology must allow for economies of scale, and steel manufacturing in the North American geography continues to be fragmented;  i.e. steel mills need to find it more cost effective to have their EAF dust recycled by Horsehead than doing it themselves.

    (c) That a new high-value use for raw EAF dust does not appear (for example, as a form of fertilizer for a genetically engineered vegetable able to extract nutrients from  EAF dust)

  2. Having the lowest-cost position compared to other horizontal competitors (zinc miners+smelters and zinc recyclers):

    (a) Through the use of the best technology, and possibly denying others access to the best technology

    and/or

    (b) By having the structural factors in place to allow lowest cost operations, which others cannot match. For example, lowest cost factor inputs into the refining process (such as energy costs), factory sites next to, or structurally well linked to minimills (e.g. dedicated and hard to replicate rail lines).


Quantitative analysis - Current and projected future cost of production
(Does the company have a lowest-cost position?)

Given that the cost of production is such a key factor in sustaining Horsehead's earnings stream, let us have a look at Horsehead's numbers, to get a "back of envelope' feel for their costs of production relative to other zinc recyclers and miners. It's likely that we don't have the calculation correct down to the penny, but that shouldn't materially affect our analysis since we are trying to figure if they have an "order of magnitude" cost advantage over their competitors.

Revenue and Operating CostZinc recycling and smeltingNickel recycling and smeltingEAF Dust Removal
Revenue$270.7m (Zinc metal: $171.2m; Zinc oxide: $96.6m)$47.7m$39.4m
Operating Cost-Total (Excl overheads, interest and depreciation)$267.3m$33.5m$24.0m
  • Operating Cost-Feedstock costs - 19% of total operating cost


  • $50.8m
(73% of raw material was EAF dust, which is "free"; 27% is zinc secondaries which is this cost)
--
  • Operating Cost-Conversion costs - 81% of total operating cost
  • $216.5m 
(cost of coke, natural gas, electricity for other sites, and coal-generated electricity for Monaca, maintenance costs, labor costs, diesel fuel for freight etc)
--
No. of short tons produced / processed (i.e.  for which revenue was received)
(2000 pounds = 1 short ton)
126,720 on a zinc contained basis
(raw tonnage: 136,661)
raw tonnage: 27,000 tonsraw tonnage: 532,000 tons
Operating cost per produced / processed pound (C1 cash cost)$1.05 per zinc contained pound--
  • Operating Feedstock (zinc secondaries) cost

  • $0.20 per zinc contained pound
--
  • Operating Feedstock (EAF dust) cost

  • $0.00 per zinc contained pound("free" - mini mills pay to have it trucked away)
--
  • Operating Conversion cost

  • $0.85 per zinc contained pound
--


Horsehead is in the process of building a new (SX-EW solvent extraction electrowinning technology) zinc refining plant in Rutherford County, North Carolina to replace its Monaca PA facility. Let's assume that when this plant goes live in Q3 2013, it will be achieve the cost efficiencies that management is likely hoping for:

(1) Feedstock costs will be zero; either because they will be able to use EAF dust as the sole feedstock, or because they are able to recover lead and silver in the EAF dust to offset the cost of zinc secondaries).

(2) a 35% reduction in per-unit conversion costs, with possibly even lower per unit costs as volume goes up, since a portion of the conversion costs are fixed. Let's also be optimistic and say that this 35% per-unit cost reduction applies across both the pre-Rutherford plant production steps that feed into the Rutherford plan and the conversion at Rutherford/Monaca (today,  CZO and Zinc Calcine are produced and fed to the Monaca plant; in future it may be Waelz Oxide and Galvanizer skimmings that are fed into the Rutherford plant).

In this best case scenario, the C1 cost of production would then be $0.55 per zinc-equivalent pound on a factory exit gate FOB basis.


Comparison with Zinc production cost across the industry

The cost of zinc mining and production varies across the world. Zinc production costs are a combination of mining and smelting costs. The final cost of zinc production is complicated as it depends on many factors. This presentation from Nyrstar gives a good overview of the cost components. For example, mining costs vary depending on the byproducts and credits that a mine can get. Zinc ores are often extracted together with Lead bearing ore, so the cost of zinc mining could in some cases be considered negative if lead ore sales cover the costs of mining. Smelting costs also differ for the ores produced by each mine, because each mine's ores are in different forms and purity. The cost to transport ores to smelters and the cost of energy for the energy intensive smelting process are also variables. For example, an impure ore that has a low zinc content will cost more to transport on a zinc-contained basis. But if a cheap source of power is available nearby, a smelter can be built that will have lower smelting costs, which could negate the cost of transporting the low grade ore. Complicating the matter is the fact that energy costs are becoming increasingly globalized (for technological and structural reasons - there are fewer cases where proximity to a source of fuel confers an energy cost advantage).

Companies such as Brook Hunt research and publish reports on estimated mining and smelting cost curves. That information is available from these research firms as a paid service. Nonetheless, there are some publicly available data points that allow us to estimate Horsehead's cost position in relation to its competitors:
  • One of the world lowest cost zinc mines is the Vendanta Resources operated Rampura Agucha mine in India. [Ref: Vedanta C1 mining costs ].  According to information provided by Vendanta, their 2010 C1 mining cash cost is $300 per ton (approx $0.15 per zinc contained pound), and the mining industry 50th percentile is $900 per ton, and the 75th percentile is $1,267 per ton.
  • Xstrata provide the overall C1 operating cost for their zinc mines, covering both extraction (mining) and conversion (smelting) costs.  [Ref: Xstrata C1 zinc mining costs].  In 2010, they report an average cost of $0.31 per zinc contained pound, and they claim that they are among the lowest cost producers.

Back-of-envelope comparisons of this nature are imprecise, but the data does give us a sense that on a factory gate FOB basis, Horsehead is not one of the lowest cost producers of zinc. Having said that, Horsehead does have the benefit of being sited in the United States. The United States imports >70% of its zinc, so the cost of transportation is a cost that Horsehead's products do not have to incur. Let's see how much the transportation premium is.

  • Based on recent data, the cost of ocean freight is approximately $0.02 per ton-mile, and the cost of rail freight is approximately $0.03 per ton mile. [ Ref: BTS study ]. The relative fuel-energy efficiency of the different modes of transport bears out this relative cost structure. [ Ref: Center for Climate and Energy Solutions ]
  • Given that the earth's circumference is around 24,000 miles, and the breadth of the continental United States is around 2,000 miles, the furthest distance imported zinc would need to  travel is likely to be 12,000 ocean miles and 1,000 rail miles. This translates into an import transportation cost of $270 per ton, or very roughly $0.14 per refined zinc contained pound (assuming high purity zinc product) ].

Which suggests that for United States customers (such as the Nucor mini-mills), Horsehead's cost position is roughly the same as a foreign producer whose FOB cost at the smelter exit gate is $0.55 - $0.14 = $0.41 per zinc contained pound. This suggests that Horsehead's cost position is comparatively advantageous. However it is a qualitative assessment whether you feel this puts Horsehead among the lowest cost producers.


This comparison does have a margin of error. Here's a couple of ways the comparison may be off base:
  • Zinc mining and production is an energy intensive one, so our cost comparison must be made against 2010 costs for the industry as a whole, so that we use the same energy cost basis.  (We must compare 2010 costs against 2010 cost curves, because energy prices change year by year.) That way our comparison will show the relative cost of production arising from the cost elements which are specific to each player (For example, comparing a 2010 cost of production with a 2007 cost of production would be misleading because the cost of carbon-based energy was different). What could trip up our comparison is if players like Xstrata have longer running energy supply contracts which expire in 2012, so their reported 2010 costs are actually based on energy prices from a few years back.

  • In future, new mines may open up which change the cost curve; for example, a new mother lode of easy to extract high purity zinc ore may be found. The probability though is that all the low hanging fruit have already been picked, and that new mines will cost more than existing mines to operate.

  • The reported C1 operating costs for mined zinc includes credits (and "profits" from recoveries of lead and other elements in the zinc ore). So we may not be looking at an apples-to-apples comparison. In reality, the cost of producing any metal (such as zinc) will vary on how much other metals are present in the ore which can also be sold. If the type of ore extracted at a mine changes over time (e.g. different strata of ore at different seams), then the cost of producing zinc at the mine will change over time.

Summary
Broadly we arrive at the following conclusions:
  1. Horsehead is not the lowest cost recycler/producer of zinc, even though its cost position is advantageous. Our back of envelope cost numbers (even with the projected cost reduction) suggest that on a smelter exit gate FOB basis, the recycling process is more expensive than the 25th percentile of  most zinc mining (extraction + smelting) costs. The question is: will this continue to be the case in future? This is a qualitative assessment that you need to make, and you may want to consider factors like:

    (a) Fundamental shifts in mining cost curves: For example, are today's low cost mines going to be exhausted in the next few years and replaced by higher cost mines?

    (b) Sensitivity of mining (extraction) cost curves to energy prices. As another example, if mine-ore extraction costs are very energy intensive, then ZINC could have a cost advantage when energy prices go up because the price of EAF dust is likely to be less energy dependent. (It is effectively free to Horesehead today) (This presupposes that mine-ore smelting and recycling conversion processes are roughly equally energy intensive; if recycling conversion is more energy intensive than mine-ore smelting, then rising energy prices may raise Horsehead's costs of production more than it does for zinc mines)

    Betting on the transportation cost advantage relative to foreign producers is a bet that:

    (c) There will not be a low cost zinc mine appearing in the United States, otherwise that zinc source could be cheaper than Horsehead; and/or

    (d) Transportation costs are not going to experience a sustained drop. If transportation costs go up, then Horsehead's competitive position will improve considerably.


  2. The viability of the ecological niche that Horsehead occupies (EAF dust recycling) depends on:

    (a) the future direction steel demand and steel manufacturing take in North America. This is another qualitative assessment that needs to be made; and/or

    (b) the propensity of mini-mills wanting to capture more the profits in the value chain. and deciding or threatening to vertically integrate


  3. The company's future earnings stream largely depends on the price of zinc. This is one of the key variables that an investor needs to project. ZINC is a reasonably good investment vehicle if you have an opinion on the future price of zinc, as it is one of the few zinc pure-plays around. However, its moderate-cost position detracts from its investment merits somewhat, as does its relatively weak balance sheet. The company may not have the financial staying power to ride out a depression in zinc prices / rise in energy prices without shareholder dilutive capital raises.


Please leave a comment or drop me an email if you have any thoughts you'd like to share.

[ Update 11 Jan 2011: Thanks to an anonymous reader who informs that the Monaca power plant has been idled. Horsehead now buys its power for the Monaca plant from First Energy (FE). ]

[ Update 4 Feb 2012: For clarity, I have edited some of the cost comparison paragraphs. Thanks to Jacob Benedict for his feedback. ]

-------------------
References:
1 ton/short ton = 2,000 pounds
1 tonne/metric ton = 2,204 pounds

Overview of Zinc Smelting and cost and profit components of zinc smelting:
http://www.nyrstar.com/investors/en/Nyr_Documents/English/zincleadsmelting.pdf

Cost curves for Zinc Mining and Smelting in 2010:
http://www.latiza.com/archivos_publicar/presentacionesconferencia_2011/Zinc%20Market%20Outlook%20-%20Lima%208%20July%202011.pdf
http://www.nyrstar.com/investors/en/Nyr_Documents/English/2011-05-12ab.pdf

Combined cost curve (mining + smelting) c 2010:
http://files.shareholder.com/downloads/HBM/0x0x446391/255a8364-6605-4296-9612-973ad9b59140/Hudbay_BMOPresentation_Feb24_FINAL.pdf

Definition of C1, C2, C3 mining costs: (Cash operating cost, C1 + depreciation and amortization, C2 + interest and indirect costs) (Industry standard nomenclature)
http://www.gfms.co.uk/brochures/precious_metals_costs_service.pdf
http://vp134.alertir.com/afw/files/press/boliden/Kokkola-2008-3_Industry%20trends_LE2.pdf
http://moneyterms.co.uk/cash_cost_mining/
http://www.nortongoldfields.com.au/PDF/ASX20April11_%20Quarterly%20Report_March%202011.pdf

Mining industry survey
http://www.pwc.com/en_GX/gx/mining/pdf/mine-2011-game-has-changed.pdf

Other EAF dust recycling companies:
http://www.globalsteeldust.com/steel_dust_recycling
http://steeldust.com/

Waelz Kiln technology:
http://www.valo-res.com/pdf/WaelzKilnDescription-EN.pdf

Overview of the Global Steel Industry:
http://www.dbresearch.de/PROD/DBR_INTERNET_EN-PROD/PROD0000000000231889/Presentation%3A+Global+steel+market+-+the+present+and+the+future.pdf

Transportation: Average Freight Revenue per ton mile:
http://www.bts.gov/publications/national_transportation_statistics/html/table_03_21.html

Transportation: Energy efficiency (BTU) per ton mile:
http://www.c2es.org/technology/factsheet/FreightTransportation
http://en.wikipedia.org/wiki/Fuel_efficiency_in_transportation


Sunday, May 29, 2011

Betting on a Housing Recovery - Title Insurance

Transaction volume in the American housing market has been in a downward trend since 2007, likely driven by 2 factors:
  1. massive over-pricing caused by speculative demand prior to 2007 - prices are gradually dropping to more affordable price-to-income ratios.

  2. a generally sour employment situation and a weak income growth, caused by a slow economy since 2008.
At some point in the future,the number of transactions in the housing market will recover because America's underlying demographic trends point to continued growth in household formation over the next few decades. Prices are likely to continue coming down as the excess housing inventory continues to be brought to market (as banks gradually write-off their repossessed properties and make them available on the market at lower prices to stimulate buyers). As prices come down, transaction volumes are like to start trending upwards once housing prices revert to more affordable levels and the employment situations recovers

One way for investors to bet on this recovery is to invest in title insurance companies. Title insurance is unique to the American market. Unlike many other countries around the world, the United States does not have a centralized register of land records. Land records, claims and liens are filed with different authorities and recorded in many different registers, depending on the jurisdiction and the laws/practices at the time of a transaction. Title insurance companies help buyers manage the risk that a title has a competing claim recorded against it somewhere. Title insurance companies do this by establishing and maintaining a database of land records, and guaranteeing to a buyer that a particular title is free from defects or competing claims, and that the buyer is able to legally take ownership of the title. They more or less are taking on the role of a national land registry in the United States.


The economics of this business are:
  1. fundamentally low risk. Unlike typical insurance companies, title insurers behave more like the operator of a reference service or a fact-checking company that guarantees its results. Unlike Life and P&C insurance companies, title insurers generally do not assume much risk in their business. Their business model is to use their databases to screen out bad risks, and only insure those title transfers which the database indicates is free from competing claims. If executed correctly, this business is one whose profit characteristics are fairly stable.

  2. not-obsoleteable, and a non-discretionary item in real estate transactions. The product is an unavoidable expense, as mortgage lenders require that title insurance be a part of every real estate transaction. The flip side of this is that broad business volumes generally follows real estate market transaction volumes. The demand for title insurance is a derived demand - it is a second order demand, which derives from the first order demand for housing. As a derived demand, there is little that management can do to stimulate demand. In this sense, its characteristics are like those of common carriers shipping lines, where one of management's primary value-add is in being able to pre-emptively adapt company capacity and costs to changes in the industry primary demand trends.

  3. moderately scalable. Once a database is built and maintained, additional business signed primarily works to reduce the per-unit cost of maintaining the database. The additional variable costs primarily come from growing and running the sales and distribution network. In downturns, these companies have to adjust by cutting the sales and distribution infrastructure. As long as business does not drop below the database maintenance costs, the business should be able to survive.

  4. highly commoditized. Apart from economies of scale, there is little competitive advantage that a firm can build over a competitor. The title insurance business is a commoditized and highly transactional one. There is little to structurally differentiate one title insurer's product offering from another. 
In the United States today, the 4 largest players are:
  • 1 Fidelity National (FNF) - 38% market share
  • 2 First American (FAF) - 27% market share
  • 3 Stewart Information (STC) - 14% market share
  • 4 Old Republic (ORI) - 11% market share
Given the economics of this business, it is highly probable that the largest player within a jurisdiction can work towards having the lowest cost, and establish a structural competitive advantage as the lowest cost provider. Because of the commoditized nature of the product, getting to the lowest cost position is a matter of who has the best execution ability and cost-focus. In other words, until an 800-pound gorilla emerges with its attendant structural competitive advantage of being the lowest-cost player, it is a tactical dog-fight and good management becomes the source of a key competitive advantage for a title insurer.  Make no mistake, this is a competitive commodity business. The single digit margins that even the largest players have suggests that competition is brutal, and product differentiation minimal.

An investor who wishes to bet on the housing recovery can try investing in either FNF or FAF. A bet on either company is fundamentally a bet:
  1. that housing transactions will start trending upward at some point over the next 5 years, benefiting all title insurers (the "rising tide raises all boats" theory).

  2. the size of the company allowing it to build in a low-cost competitive advantage over smaller competitors.

  3. the management of the company being able to out-execute its competitors, and building it into the lowest-cost player
Khrom Capital has identified another title insurer (ITIC) which it believes is a good investment opportunity. Read about it here on Scribd: Khrom Capital on ITIC .  

What can go wrong for title insurance companies?
  1. Government regulation can become onerous and impose pricing controls on title insurance products.

  2. The United States could start implementing a system of land registration, which would obviate the need for title insurance.

  3. The United States banking industry is becoming more concentrated, with the top 5 lenders accounting for an ever growing share of mortgages in the market. It is potentially detrimental to a title insurance company/industry if they decide to come up with their own title insurance system, or boycott any of the title insurance companies.
     
  4. Title insurance is offered "in perpetuity" to a buyer, and there is typically no limit on when a claim may be filed. For example, a title insurance company is on the hook if there is a contest to a title even 15 years after the purchase of the property, as long as the buyer from that purchase still has a stake in the house. Any defects in the title records / title plants of the company today may not be known till many years down the road.
[ Update 29 Jan 2012: An excellent introduction to title insurance can be found at this website:  http://www.mtgprofessor.com/a%20-%20title%20insurance/questions%20about%20title%20insurance.htm ]

    Sunday, April 24, 2011

    Dominant firms and Wide moats : How moats are breached


    Incumbent dominant firms typically have a strong competitive position. These companies are dominant because they have out-executed the competition, and have found the most successful playbook for their ecosystem. If the firm's dominance came after years of open competition, then chances are that the winning playbook has been battle tested against almost every other type of competitive strategy. In other words, you'll be hard pressed to find a better way to win in the ecosystem. These firms have identified (consciously or unconsciously) what their customers value, and have built their operations structure (systems, social proofs, measurements, infrastructure) to delivering to that need in the most effective way.

    For example, a dominant FMCG firm is likely to have found the best mix of how to elicit customer needs, carry out chemical R&D, design and manufacture, sell through retail channels, and build mind share in the consumer consciousness. A dominant retailer is likely to have found the winning formula for selling to a particular consumer segment, delivering the desired products at the desired price to consumers at the time and place that they prefer. These companies are likely to have shaped themselves to exploit any and every edge and efficiency in the business, to dominate over their competitors.

    Some industries are not amenable to concentration or lack structural economic effects, for example: the business of mining and supplying construction aggregate, flour milling and cement manufacturing. In these industries, the dominant firm is likely to be an extraordinarily canny executor. In industries where there is an economic effect (for example, economies of scale or the network effect), the incumbent will have an even wider-moat around its business.

    Investors often bet that these companies will be able to keep the competition at bay, and provide a stream of earnings that can be estimated with a higher degree of certainty. So one of the key risks that these investors need to assess is the probability that the company's moat will be breached. In other words, "What Can Go Wrong?" or "How might a competitor breach the moat and dethrone the king?"


    Dethroning a dominant firm in an ecosystem is very difficult. A look at history and the world of business suggests that a competitor can only make headway if (a) the management screws up and destroys the business from within, or (b) some facet of the ecosystem has changed and the incumbent has not recognized or reacted to it. The former is entirely a function of how competent management is, and as investors, the question is whether shareholders will be able to effect management change if the executive office goes sour. Management can go sour when incompetency sets in, or when management simply forgets the playbook that made the company successful.

    Ecosystem changes on the other hand can cause an irreversible decline in a company's prospects under certain conditions. A business ecosystem can be broadly thought of as the combination of: (a) the nature of source of demand, (b) the nature and source of supply, (c) the tools, regulations, technologies and factors of production which companies use to create value for customers. Ecosystems typically change in one of the following ways:

    (1) a disruptive ecosystem changes occur (such as the internet changing the playing field for newspapers, or the change in zeitgeist from one fashion trend to another). These typically cause the incumbent and the industry to go into terminal decline. Mature established businesses rarely have the organization tools, social proof mechanisms and culture to allow them to operate like startups in a greenfield area. Further, the organizational imperative and social proof will drive management to defend the established cash-cow at all costs, which causes the organization to dig itself into the dying ecosystem further.

    (2) the ecosystem experiences incremental changes that change the shape, but not the essence, of how businesses operate in the ecosystem. These changes typically come in the following forms:
    1. A new way of running operations (for example, brought about by technology, new thinking, or infrastructure changes) appears. In this case, a competitor can leverage this new approach to become better at the game than the incumbent. For example, Walmart took the approach of going after 2nd tier towns and building geographically contiguous logistics facilities, and achieved a lower cost of operations than K-Mart. By failing to take on this new operating approach, K-Mart allowed upstart Walmart to grow and eventually dethrone it.

      Dominant firms need to be constantly adopting incremental changes to keep operations and its business at the forefront, and to minimize any comparative efficiencies that competitors can exploit to operate more effectively than the company.

    2. A new market segment starts to become viable.
      (a) For example, the rising incomes in China and India make it imperative for dominant firms like KO and PEP to start selling there even if profits are not apparent. Apart from the need to build a future consumer base, allowing someone else to serve these large new emerging markets first would sow the seeds for a big competitor to come up in future, and dethrone the incumbents. (The corollary is that the maxim "China is market you have to be in" is really only true for large dominant firms that need to defend their positions. If you are a small enterprise, plunging into a big foreign market isn't something that needs to be front-and-center on your radar)

      (b) as another example, major banks in the United States are trying to serve the market of "the un-banked" blue collar immigrant workers. Failing to do this may allow lateral players who have relationships with these consumer, such as WU, WMT, cash advances and pawnshop operators to build a firm foothold which could allow them to gradually expand and one day challenge the banks in the financial services arena.

    3. Not responding the shifts in consumer demand or zeitgeist. A dominant firm that fails to read its consumers can cede ground to competitors. For example, for quite sometime Coca-Cola focused primarily on its flagship cola, even as consumer tastes gradually become more disparate. If KO had not reacted in time and diversified its product line, it is entirely likely that a competitor could have grown large enough to challenge KO simply by selling the sodas, still waters, teas and other drinks that the public wanted. If that happened, one pillar of KO's moat - it's dominant distribution system, would have been severely compromised.

    4. A new form of supply emerges. A dominant firm fails to respond to changes in sources or nature of supply. For example, consider how Les Schwab managed to compete against the dominant tire-store chains (who were tied to then dominant American tire manufacturers) by selling Japanese made tires. (An equally intriguing question is how did the Japanese tire manufacturers compete against the dominant American tire manufacturers? The answer is that they rode the wave of Japanese car exports. Which begs the question of how did Japanese car manufacturers manage to invade the home turf of the US automobile giants? They likely made headway by leveraging the oil crisis shock to the automobile-industry-ecosystem, which caused consumers to start looking for more fuel efficient cars. Japanese manufacturers wedged in through that shock and then continuously out-executed the competition through innovation and product quality)
    For a competitor to exploit these, speed is of the essence. The competitor must quickly build itself up before the dominant firm realizes what is happening. Dominant firms are typically very well placed to close these gaps because:
    1. the incumbent has an in-place infrastructure and cash flow, giving it a whole lot more resources than the upstart competitor. (In many cases once a dominant firms starts addressing a gap that a competitor is trying to fill, the economically sensible outcome is for the competitor to sell itself to the dominant firm)

    2. the incumbent has customer share of mind; Customers know the incumbent firm, and the inertia of customer behavior allows the incumbent to swing the customer back to its offering once it closes the gap with the competitor.

    Because of the advantages of incumbency, It is often the case that incumbents (apart from situations of disruptive ecosystem change) cede ground to competitors only when (a) an ecosystem change occurs, and (b) management fails to reach because of hubris, inattention, or other reasons, and (c) the new competitor has extremely good execution capability.

    For investors, the takeaway is:
    (1) at the right price, buying the incumbent dominant in an ecosystem is a good bet
    (2) but you must assess the parameters of the industry ecosystem in which the firm operates, and watch for changes to it

    * In addition to the other things you would need to assess for any investment. For example, factor additional risks for industries where non-economic actors are likely, for example strategic industries prone to government intervention (e.g. Petrobras, power utilities), or vanity industries like sports teams where owners are looking for "fun" and not profits - in both examples, competitors can operate at a loss for long periods of time and destroy any dominance an incumbent may have.

    Sunday, March 6, 2011

    Microsoft (MSFT) Investment Thesis / Stock Analysis


    Microsoft is one of the blue-chip companies that are trading today at extremely attractive prices. By any measure, Microsoft is an impressive business machine; the economics of its business allow it to achieve ROAs in excess of 20%, ROE's in excess of 40%, and net margins in excess of 20%. To top if off, Microsoft has managed to grow its revenue and earnings per share at a CAGR > 10% over the last ten years. By the numbers, this looks like a growing company with excellent financial characteristics.

    Yet its share price has gone nowhere over the last ten years. With its EPS increasing over the years, its PE ratio has dropped to approximately 11 times 2011 forward earnings.

    This stunning dichotomy simply demands a closer look by any investor worth his salt. This is either an unprecedented buying opportunity, or Microsoft's business is about to enter a rapid descent, which is what its current valuation suggests the herd is thinking. The latter also demands the attention of any serious investor, as a case study of how a wide moat business can lose its edge and fall into the abyss in a blink of an eye.


    Microsoft's earnings streams

    Microsoft is really a combination 4 separate earnings streams: (1) its profitable and well established Windows Operating System and Microsoft Office earnings stream, and (2) earnings from its server software division, where it is a competent player, and (3) earnings from its online and games division, a new area which the company is trying to break into, and (4) its nascent Azure application development cloud its Office and Exchange cloud businesses.

    The bulk of Microsoft's earnings come from its Windows and Office product lines. These earnings stem from its dominance of the operating system and office productive software markets for the PC (and PC server) space. They make up the bulk of its earnings and revenue, and enjoy enviable margins.

    -----------------------------------------------------------------------------
    Sidebar: Here's how margins(net income/revenue) stack up in the software industry:
    NOVL: 0% CSC: 5% RHAT: 12% SAP:15% CA: 10-16%
    ADBE: 20-22% ORCL: 22-24% MSFT: 27-29%
    MSFT Operating income/Revenue: Server and Tools = 33%
    MSFT Operating income/Revenue: Business Division = 61%
    MSFT Operating income/Revenue: Windows And Windows Live = 67%
    -----------------------------------------------------------------------------

    Valuing Microsoft broadly boils down to thing things: (1) valuing its established Windows and Office franchise, and (2) valuing its cloud, gaming and online initiatives. The first requires an analysis of the moat and sustainability of its PC based ecological niche, while the latter requires a Venture Capital type assessment of a startup's prospects. Its server software business, while competent and well executed, is not large enough to move the needle on valuing the Microsoft behemoth.

    Assessing Microsoft's prospects in its new initiatives is a challenging task. As any business historian can attest, the track record of established companies successfully remaking themselves in new ecological niches is dismal. It is exceedingly difficult for established companies to innovate and enter new market spaces. A look at the history of capitalism suggests that groundbreaking innovations are usually brought to market success by startups. Established businesses have rarely been able to create, innovate and enter fundamentally new market spaces. Where they have done so, it's typically by acquiring upcoming startups.


    How businesses evolve in Ecological niches, and the challenges dominant firms face in evolving with changes in the Ecology (Business Anthropology)

    Most established companies are the survivors of the ecological market niche that a particular confluence of technology, fashion or the zeitgeist that happened at point in history. For example, P&G and Unilever are the dominant survivors of the ecological niche brought into existence by the advent of chemical manufacturing, mass communications, and logistics. Similarly, Coca Cola, Kraft, Nestle and Heinz are the dominant survivors of the intersection of mass manufacturing, agricultural logistics and mass communications.

    The same is true in the technology industry, except that ecological niches exist for much shorter time spans. IBM was the dominant survivor of the mainframe era, but that ecological space shrank precipitously in the 70s. Digital Equipment Corporation dominated in the minicomputer space, an ecological niche which all but disappeared by the 90s with the coming of the microcomputer ecology. Microsoft and Intel are in turn, the dominant survivors of the microcomputer ecological niche.

    Few large companies manage to recreate their success in new ecological spaces when their original ecological spaces expire. Polaroid and Kodak declined when their spaces declined; Newspapers, print media (and likely cable companies) are starting their decline as their ecological space is eroded by digital technology. Those that succeed in surviving the decline of their niche typically lose their dominant franchises. IBM managed to plant itself in the new ecological niche of IT services, but it no longer enjoys the franchise economics that it enjoyed during the OS/360 mainframe era. It's economics today are more of a competent well-executed business, as opposed to one enjoying franchise economics.

    There are many reasons why large companies fail to do this. Clayton Christensen's The Innovators Dilemma presents one reason. I believe the primary reason is that what made them large dominant companies also makes them bad at the core innovation and flexibility needed to experiment and succeed in a new ecology. In their nascent stages, these companies were nimble startups that worked out the core value proposition and play books for their ecological niches. In any new industry in a competitive economy, thousands of startups will try their hand. Only a few find the sweet spot which consumers are willing to pay for. Of those, the dominant survivors are those who can execute the business best; with the best planning, resource allocation, logistics, and discipline to allow it to marshal resources and deliver the value best to consumers. They out sold, out delivered, and generally out planned their competition. These companies have built their measurement systems and social proof by these lines, allowing them to execute a business play book successfully.

    Unfortunately this is also the type of environment that is the complete opposite of a startup. As a result, fundamental innovation is rare in dominant firms. The employees and the firm primarily excel and are measured at executing the delivery of a value proposition, a value proposition that was defined early on in the company's history and around which the entire business' hard and soft infrastructure is constructed. This isn't to say that they are completely devoid of innovation; often they excel at incremental innovation (for example Toyota), or have established systems to acquire or compete companies in in adjacent ecologies where other pathfinders have established that a market exists (for example Nestle).

    Microsoft seems to exhibit the same characteristics of large dominant companies. In spite of Microsoft's enormous war chest and extremely farsighted (some would say paranoid) management who have demonstrated more vision and nimble thinking that many, its forays into the online, gaming, and non-PC niches have not been spectacular.
    • The four horseman of the Internet (Ebay, Amazon, Google and Facebook) have firmly established dominant positions in the recently evolved Internet ecological niche.

    • In the emerging mobile form-factor ecological niche (PDAs, mobile phones, tablets) Google and Apple have run away with the lead.

    • It is an open question whether Microsoft's cloud gambit will pan out. Already Google, Amazon and Force.com have already made forays into the productivity and applications hosting cloud platforms, albeit with different value propositions. (The applications cloud space is not a new one, companies like ADP have been hosting applications for decades - Before "Cloud Computing" became the buzzword of the day, applications clouds vendors were known as Application Service Providers. What has changed is that there are now a much larger range of applications which are now being delivered in this way; partly because technology now allows for it, and partly because a critical mass of software use-cases have evolved to a steady state, allowing companies to make the large capital investments needed to build cloud-delivered versions of the software without worrying so much about whether the features built will be accepted).

    Valuing Microsoft

    Given the dismal track record of dominant companies surviving the demise of their ecological niches, it would not be prudent when valuing MSFT, to bet that the company will be able to transform itself and dominate an upcoming ecological space. Rather, the question when valuing MSFT is:
    1. whether the PC space will continue to coexist with new ecological spaces, or whether we are at the transition between the eclipse of one space and the rise of another, and

    2. the source of MSFT's strength, and what can destroy MSFT's moat in the PC space.


    Sustainability of the PC ecological space

    The computing industry has been progressively shrinking the size of computers, from the mainframes to todays' PCs and mobile computers (PDAs, cell phones, netbooks). From requiring teams of people to operate a mainframe system, to allowing individuals to own and use computing. It has been about a continuous reduction in the form factor of a Turing computing unit, and removing the accidental difficulties in programming and software.

    The question now is whether the PC (desktop or netbook) species is in a punctuated equilibrium within the computing ecological space - in other words, whether it is a form factor that will continue to coexist with whatever else may come along. I suspect the answer is yes. It is likely that the PC form factor is the most suitable form factor for many types of Information-based work built around the keyboard-monitor combination. This is a spectrum of work that requires textual input and concentration on the display, and the PC (desktop or laptop) is the culmination of the various evolutionary experiments of what type of machine form factor is best suited to it. The PC is likely to become another office fixture or household appliance, like a household refrigerator that coexists with wine coolers and other appliances.

    This doesn't mean that the market size for this form factor will remain what it is today:
    • There are likely to be many people who are using this form factor today who would be better served by something else that comes along. These are likely to be persons whose work with the computer is less intense, or is primarily information consuming in nature. Many people who typically just browse and surf will switch away from PCs to tablets and other information consumption devices. They are now using PCs simply because there is no better option.

    • As the desktop computing usage pattern enters punctuated equilibrium, we may see the desktop replacement cycle approaching the once every 10-years replacement cycle of refrigerators - people only replace them when they break down.
    Unless a larger proportion of humanity takes on desktop computing, it is entirely possible that the desktop PC market will shrink until it reaches its natural stable size.


    Microsoft's moat in the PC ecological niche

    The question then becomes how sustainable Microsoft's Windows and Office position is in the PC space. Windows and Office's wide moat is primarily because these 2 types of software have Platform Economics, and Microsoft has a play-book that successfully made the most of this economic effect. A secondary factor is that Microsoft has a well-run and well-staffed software development organization that has been able to continually respond to changing customer needs, by adding new features/releases and supporting its existing software.

    Platform economics is a term I use to describe an economic effect seen in certain industries. A product enjoys platform economics when it has gravitational mass. As its gravitational mass grows, it attracts more matter to it, causing it grow even more. In a platform, this gravitational mass is conferred by the product enjoying 2 or more of the following economic effects:
    1. it enjoys a network effect, and
    2. it enjoys economies of scale in a positive feedback loop, and
    3. it guarantees an optimal solution to the "prisoner's dilemma" or "tragedy of the unregulated commons" for all persons using the product
    The Microsoft Windows operating system enjoys this economic effect. In the era before interoperability technologies and standards were common, it was the case that as more people used Windows, each user benefited because he/she was able to exchange files with an ever larger community of people. The larger community also made it more cost-effective for software vendors to write software for the Windows operating system, increasing the number of applications available to Windows customers, further increasing the value of the operating system to its existing and potential customers.

    The Windows operating system also provided software programmers and hardware manufacturers the promise of breaking the prisoner's dilemma. In the early days of desktop computing there was a proliferation of hardware devices and software programming platforms. Individual software vendors had to write their software to support the multitude of hardware devices out there. Software programs like WordPerfect 5.1 distinguished themselves by having the code to support a large number of printer hardware devices. Printer manufacturers like Epson distinguished themselves by canvassing as many software vendors as possible to support their printers. While this was an environment that worked for each party, it was far from optimal. (The system had settled into local optima instead of the global optima) Microsoft played the role of a global optima guarantor through the Windows operating system. It created a common device programming model, so that all software vendors only had to write to support the (one) Microsoft programming model, and all hardware vendors similarly only had to support the (one) Microsoft device driver model. All participants in this scheme stood to benefit, but only if everyone participated. A software vendor could support all possible hardware versions, and vice versa. Microsoft succeeded in delivering on this value, and Windows became the prisoner's dilemma global optima guarantor.

    To a lesser extent, the Microsoft Office software also enjoyed platform economics. As a productivity software suite, it's users needed to exchange its files with other users (exchanging Word documents, Excel spreadsheets and so on). The more people who used Office, the larger the community of people its customers could exchange files with. The more people using it also made it more attractive for software developers to build productivity add-ons like macros and formulae in its Excel software and templates and fonts for its Word software.

    Microsoft's playbook, which integrated its software products and allowed them to exchange information with each other, as well as its cultivation of the software developer community, fully exploited the platform economics characteristics of its products and created a wide moat. Customers who bought Microsoft products entered an ecosystem of benefits that only the Microsoft family of products could provide.

    Supporting Microsoft's play-book was a solid software development team. There are 2 unique characteristics of software that make it different from analyzing other industries:
    1. Software is all about design. The source of competitive advantage is in having a business model that supports a team of developers who can keep updating software to keep pace with changes in the way customers work, or the way information is being processed. Constantly designing and redesigning software to meet changing usage patterns and needs is the key to sustaining a competitive advantage. There is no manufacturing economies of scale in software, because software is an information good with zero marginal cost of production.

    2. The way to get an edge in software development is to hire the best people. As primarily a design activity, hiring the best developers in the industry brings you much closer to having the best software development capabilities.
    On both counts, Microsoft was firing on all cylinders throughout the 90s and early 2000s. The early PC era was like the Cambrian explosion, as people tried many different ways of processing information, use-cases, usage patterns and different types of software. Microsoft was able to build an organization that supported constant software adaptation to new and evolving usage patterns and needs. It was also seen as the most attractive place to work for bright software engineers, and it had a business model to fund its software development talent and organization. It built a market-to-lab feedback loop that allowed it to come up with continuous software design changes (upgrades and releases) to meet evolving customer needs.

    Threats to Microsoft's moat

    As computing evolves, we are now at a juncture which presents the following threats to Microsoft's moat and earnings stream:
    1. Software Cambrian explosion tapering off. The evolution of software use-cases and PC usage patterns are reaching a punctuated equilibrium. This is readily observable to both practitioners and customers. Think about how much Word Processing has changed for you in the last 2 years? Not likely to be much. For practitioners, how much has relational database or core operating system technology changed over the last 2 years? This relative stasis has the following effects:

      (a) It makes it possible for the bazaar model of software development to become practical. (See The Cathedral and the Bazaar). When software patterns stabilize, it is less of a necessity to have a paid team of developers constantly tinkering with the design to meet rapidly changing needs. Because software is all about design and has zero marginal cost of production, it allows startups and loose coalitions of developers to also develop competing software. The evidence for this is in the encroachment of open source software development in areas like operating systems, databases and web servers, where software evolution stasis is setting in. This is a threat to all software businesses, including Microsoft. It is especially threatening to Microsoft because operating systems and productivity software appear to be reaching punctuated equilibrium. What works in Microsoft's favor on the desktop side is that there is still some evolution in the types of hardware devices used and end-user programs, which keeps Linux and open-source competitors at bay. The bazaar model is more suited to point, less dynamic, software which changes less and has arrived at some emergent pattern. In comparison, the bazaar model likely to hit Office first, because it is more of a point solution and doesn't have to deal with as many integration and complexity variation points. The productivity suite space (Microsoft Office) is the canary in the coal mine to watch for the bazaar approach'es encroachment onto Microsoft's moat.

      (b) Punctuated equilibrium also allows open standards to come into existence. During periods of rapid software design evolution, it is difficult for open (interchange) standards to be maintained because too many parties need to sign-off on each change. But when categories of software reach evolutionary stasis, it becomes far more likely that open standards can be negotiated and maintained. Upstart competitors are likely to be strong advocates of standards, since it gives them the means to enter the market against Microsoft. Open (interchange) standards have the effect of breaking the closed ecosystem and platform economics of Microsoft software. With open standards, you no longer need to buy Microsoft software to exchange information with other people. The computing community as a whole enjoys the network effect as more users compute, instead of a world where network effect only exists within software platforms (Microsoft's or otherwise).

      (c) It reduces the value of software having development platforms for 3rd party software plug-ins, which is a core driver of platform economics. There is less value to Windows or Office providing an platform for 3rd parties to build applications on top of it. In the old world, more developers building more applications increased the value of the platform. But in a punctuated equilibrium, a single vendor can identify what is needed by the bulk of users and build it in an integrated stack/device. For example, Apple is doing that with its consumer devices - it has identified the core applications and use-cases for the bulk of users, and built it into one well executed device. This reduces the value of having a platform like Windows or Office.


    2. Interchange standards are becoming entrenched, fueled by the Internet and the Proliferation of non-PC devices. Interchange standards remove a key pillar of the platform economics that Microsoft has enjoyed to date. If the world were such that Microsoft software existed on all types of computing devices, then the defacto standard would be Microsoft's and Microsoft would still enjoy the benefits of platform economics. However, this isn't the case, and the probability of Microsoft reasserting itself across all computing devices is low. The proliferation of other non-windows devices accessing information (over the net) accelerates the adoption of standards and provides the environment and incentives for all players to work out mechanisms to share data and standardize formats. The cat has been let out of the bag, and there is no going back.

      Not that Microsoft didn't try. Bill Gates famously recognized the threat of the Internet very early on in 1995, as seen in his "Internet Tidal Wave" memo to employees. But even with this prescience, Microsoft failed to fend off the threat. It tried shaping the Internet conversation with IE, for example by only allowing IE to access hotmail, and getting website to use ActiveX controls which only PCs with Windows would be able to access. If they had succeeded, we could today be conceivably living in a world where all websites and browsers needed ActiveX technology and by proxy windows clients. With 20/20 hindsight Microsoft could have won the browser battle, but only if they had run it lean and mean. It could have won in this new ecology albeit with less of moat, because this ecology doesn't allow it. But it was playing by the play-book of the PC era ecology, which required complete domination. Its browser grew bloated, and mis-steps on standards (by trying to freeze other players/java/etc out too early ) and fundamentally failing to create a good user experience lost it the war. Firefox and Chrome took market share because they out executed and were leaner, and Internet standards are well entrenched today.

      Microsoft also failed to create compelling experiences on tablets, PDAs and Smart phones. It is likely it failed because it was playing by the PC playbook, and didn't recognize that a new user interface and usage model was at play. Apple and Google realized this when they looked at the market with a fresh pair of eyes. Apple and Google out competed by creating compelling user experiences that were suited to the new form factor.

      The proliferation of other devices and standards also presents a negative feedback loop for Microsoft. For example, as more people buy iPhones, they may decide to buy Apple computers to sync up with those phones. Because of the punctuated equilibrium and interchange standards, they may find they are able to do most of their work on the Apple computers. This would further reduce the moat Microsoft has around its Windows operating system.


    3. Technology evolution is shifting the battle has shifted to a new arena. As the industry evolves away the accidental difficulties in computing, Cloud Computing has become a viable architecture and software delivery model. Cloud Computing, more accurately defined as "presentation layer on the client, everything else running on a server somewhere else" has reduced the raison d'etre for the various features in PC operating systems like Windows.

      It also fundamentally changes the battlefield from the PC operating system and PC applications to the applications tier (which is increasingly residing on the server). It has changed the arena "playing field" where battles are fought from the PCs in our home and office to the servers and applications running on them. In a Cloud Architecture, the ecosystem is less important on the client side, especially if you only use cloud applications. However platform economics exist on the server application side. In this new arena, the battle is to get as many users onto your cloud application as possible, so that they can enjoy the interchange and platform economics amongst themselves. Microsoft will have to try to get the bulk of its users onto the web version of office, while trying to keep Google Apps at bay. Platform and Information economics work to benefit and build a moat around the dominant Cloud platform. So Microsoft's Azure and Office 365 cloud initiatives are critical plays for Microsoft.

      The paradigm shift from PC era computing to Cloud Architecture cannot be understated. It is akin to the point where people change from installing their own power generators to using centrally generated power. But we should not overstate the case either, it is unlikely that all software is suitable for a Cloud Architecture. But for users who only need applications which are Cloud-able, the value of the PC ecosystem to them is greatly diminished.

      This type of major technology disruption is what allows upstarts to destroy dominant firms. It exposes a new front for a competitor to come in, much like KMart's failure to serve second tier cities allow WalMart to build up its momentum and eventually displace it. As a new software architecture and delivery mechanism, it erodes the need to install software on PCs. It may allow new innovations like Google's Chrome based laptop to take share, and allow Force.com and Google Apps to take share.


    Summary

    It is likely that the ecological niche which sustains Microsoft will continue to exist for some time (a decade or more). It is probably an exaggeration to say that all software will shift to a Cloud Architecture, or that PCs will disappear from the face of the earth. It is also probably an exaggeration to say that Open Source software will destroy all commercial software companies. For many types of software, software will never fully stabilize into a punctuated equilibrium, because the software is built around how humans work in the real world. And that always changes. For example, in Enterprise Resource Planning and Accounting software will have to constantly evolve to meet changes in the business and regulatory environment.

    On balance, a valuation of Microsoft relies on two things:
    1. your assessment of the qualitative issues discussed here, and how much they will (a) reduce the size of PC ecological space, and (b) reduce Microsoft's pricing power; and

    2. your assessment of Microsoft's prospects with its Cloud Computing efforts. That is the emerging threat and new ecology in which it appears to still be in a position to set the agenda.

    By any measure, this is not an easy investment case to work out; there are many elements which require Venture Capital type thinking. Here are some other investors' thoughts on Microsoft as an investment: