Capturing value from innovation: Teece's three questions for founders

A founding team has just finished its product. They know a machine shop or a large retailer could turn out a near copy within months. The question left is who keeps the profit. Teece's (1986) profiting-from-innovation framework rests on three building blocks, which I phrase as three questions. Unlike the post on innovation versus invention, which is about naming an innovation correctly, this one is about who gets paid once the product works.

Why is the first innovator not always the one who profits?

Because profit follows things other than the technology. Teece aims at science- and engineering-driven firms that believe a product meeting customer needs guarantees success. It may do so "for the product, but not for the innovator" (Teece, 1986, p. 285).

EMI is the classic case in the paper. Its CT scanner needed a high level of training, support and servicing. By Teece's account EMI had none of those capabilities, could not easily contract for them and was slow to see how much they mattered; it probably could have partnered with a company such as Siemens. Legal protection was also thin, because once the machine was on the market it could be reverse engineered (Teece, 1986, p. 298). Within 6 years of entering the US in 1973, EMI had lost market leadership (Teece, 1986, p. 286). IBM went the other way: its 1981 personal computer was not considered advanced in architecture or components, yet it shipped over 500,000 machines by the end of 1984 (Teece, 1986, p. 286), and Teece calls the reputation behind IBM's name perhaps the greatest cospecialised asset the company possesses (Teece, 1986, p. 299). This is Teece's reading; I have not checked it against independent histories.

The library's note on the framework names the usual mistake: discussing only the technology and forgetting complementary assets.

Question 1: who controls the complementary assets?

Whoever holds the distribution, manufacturing, after-sales service and reputation the product needs to reach the market. Almost every commercialisation needs marketing, competitive manufacturing and after-sales support (Teece, 1986, p. 288). There are three kinds of complementary asset.

Generic equipment and skills are almost always on hand. Specialised assets are hard to buy by contract, because the party making the dedicated investment carries serious risk (Teece, 1986, pp. 291-292). A founder can draw four rows (distribution, manufacturing, after-sales, reputation) and ask of each: generic or specialised, and who holds it now? To test whether an asset you plan to build is worth the money, use VRIO, and see why a negative VRIO result is still useful.

Question 2: is the appropriability regime tight or weak?

The appropriability regime is the set of environmental factors, apart from firm and market structure, that govern an innovator's ability to capture profit: the nature of the technology and the efficacy of legal protection (Teece, 1986, p. 287).

Patents seldom give perfect protection: many can be "invented around" at modest cost, and they are especially weak for process innovations. Trade secrets work only if the product can reach the public while the technology stays hidden, which usually means formulas and processes. Tacit knowledge is hard to transfer unless its holders can demonstrate it (Teece, 1986, p. 287). Teece divides the world crudely into tight (the Coca-Cola formula) and weak (the Simplex algorithm), admits this is a "gross simplification", and says tight appropriability is the exception rather than the rule (Teece, 1986, pp. 287, 290).

A practical question, mine and not Teece's: if a rival took your product apart, what could it copy within a quarter?

Question 3: has a dominant design emerged yet?

A dominant design emerges when, after trial and error in the market, one design begins to look the most promising and meets a whole set of user needs fairly completely. Teece's examples are the Model T Ford, the IBM 360 and the Douglas DC-3. After that, competition shifts towards price (Teece, 1986, p. 288).

Before that point complementary assets matter less, because the race is to find the winning design. Afterwards, access to them becomes decisive (Teece, 1986, p. 291). The risk for the innovator is that followers modify the product before the dominant design appears, and the modified version may be chosen as the standard (Teece, 1986, p. 288). Where protection is weak, Teece advises letting the basic design "float" until the evidence points to the likely standard, and staying closely coupled to the market (Teece, 1986, pp. 290-291).

How do the three answers lead to integrating, collaborating or licensing?

By the combination of the first two answers. The third only changes when those assets become decisive.

Each route has a price. Contracting saves capital, but a partner may become a competitor (Teece, 1986, pp. 293-294). Integration brings control, but building a specialised asset is pointless if imitators can build it faster and cheaper (Teece, 1986, p. 295). Contracting for a critical capability when you should have built it can nurture an imitator better able to serve the market than you are (Teece, 1986, pp. 297-298).

What should Mr Tam do about an easily copied water filter?

A hypothetical example: Mr Tam is a character I made up, and every figure below is invented for illustration. He makes an under-sink water filter. The housing can be copied on sight. What is special is the filter cartridge, built to a material formula he tunes himself, and customers replace it twice a year.

The cartridge row shows up in the profit arithmetic. One filter retails at VND 2,400,000, the chain keeps 25 per cent (VND 600,000), and unit cost is VND 1,350,000, so gross profit is VND 450,000. A cartridge sells for VND 300,000 online; after 20 per cent marketplace and delivery fees that leaves VND 240,000, against a cost of VND 90,000, so VND 150,000 profit each. Over three years a customer replaces 6 cartridges: VND 900,000 of cartridge profit, double the profit on the machine, and VND 1,350,000 in total, of which cartridges are 67 per cent.

Suppose another workshop launches a compatible cartridge and takes 60 per cent of replacements, so Mr Tam keeps 40 per cent. Cartridge profit falls to VND 360,000, total profit to VND 810,000, cartridges are 44 per cent, and total profit is down 40 per cent. The protection of the cartridge, not of the housing, decides most of the profit. The classifications in the table are my judgement for the hypothetical, not conclusions from the source.

When does the framework fit, and when does it not?

It fits when the product is easy to copy, needs capabilities downstream such as distribution or service, and you are smaller than whoever holds them: large firms are more likely to hold specialised assets already, small ones less so (Teece, 1986, p. 301). It does not yet fit when you are unsure there is a market, because the analysis assumes the innovation has one (Teece, 1986, p. 300).

Going deeper: where does the framework come from, and what is missing?

The origin is Teece (1986) in Research Policy, volume 15, issue 6, pages 285–305 (Teece, 1986). I read a full-text copy hosted on a third-party site, not the publisher's. The text extraction has recognition errors and a few cut-off paragraphs. Figures 1–11, including the two decision charts in figures 10 and 11, could not be read, so I rely on the prose that describes them. Page numbers follow the journal's pagination.

Teece states some limits himself: the tight-weak split is a gross simplification (Teece, 1986, p. 287); the Abernathy and Utterback dominant-design model suits mass markets better than small niches (Teece, 1986, p. 288); and the analysis assumes a market exists (Teece, 1986, p. 300).

Jacobides, Knudsen and Augier (2006) appeared in Research Policy, volume 35, issue 8, pages 1200–1221 (Jacobides et al., 2006). I read only the abstract, from the SSRN working-paper version; the full texts on SSRN, ScienceDirect and ResearchGate all refused access when I tried. According to the abstract, the paper extends Teece by pointing to the importance of "industry architectures", sector-wide templates that circumscribe the division of labour. The authors argue firms can create an "architectural advantage", high value appropriation without vertical integration, by enhancing both complementarity and mobility in the parts of the value chain where they are not active (Jacobides et al., 2006). My reading, from the abstract alone: who captures value depends on position in the industry architecture, not only on your own assets.

The framework does not say what to charge or how to raise money, and it cannot measure how tight protection is, so every rule above is qualitative. It does not cover intellectual property law in Vietnam. I have not read work after 1986, so I do not claim how the framework handles digital platforms.