ROI: The Innovation Killer

Why is it that established companies seem to have so much trouble with responding to (or developing their own) innovative products and services? Why have so many companies, both big and small, seemingly given up on solving big problems, and are instead happy to make incremental improvements to existing products? One big reason, in my opinion, is the focus in American business on Return on Investment, or ROI. Return on Investment is a measure of how much money, measured as a percentage, a given investment will return over time. The money comes from cost savings, increased sales, or both. Companies evaluate investments against a target Rate of Return that they set. If a proposed investment is expected to meet or exceed the company's Rate of Return, the company makes the investment; otherwise, it doesn't.

It's fairly straightforward to calculate the expected ROI for a new product or capital investment that's very similar to products that you (or others) already sell, or capital investments that you (or others) have already made. If you're Proctor & Gamble and you're considering introducing a new flavor of Crest toothpaste, you have a huge database of historical information about how much it cost to develop new toothpaste flavors, put them into production and market them, and how well they sold over time. If you're Amazon and you're considering building three new warehouses, you know with great accuracy how much it cost to build similar warehouses in the past and how long it took to break even on those investments.

On the other hand, consider what IBM had to deal with when it decided to launch its own personal computer in 1981. Its experience with building mainframe computers was useless in projecting the costs of developing, manufacturing and marketing a personal computer. The personal computers that had sold to date were intended for hobbyists, which was a very small market. IBM had very little information with which to calculate the ROI for its personal computer.

The more innovative a product is, the more it represents a discontinuity (a break from previous technologies, goods and services,) the less reliable are its Return on Investment calculations. To reduce risk, most companies will assign such investments a relatively low ROI. Then, when they compare the ROI with the company's target Rate of Return, they'll kill the project if the ROI is below the Rate of Return. Company managers can guarantee that the project will be killed by deliberately putting conditions on the ROI calculation that will force it to be below the Rate of Return.

Big companies innovate by accepting projects that have risky ROI calculations, or by ignoring the ROI analysis altogether. Bell Labs was able to invent the transistor because there was no requirement that John Bardeen, Walter Brattain and William Shockley work on projects that would generate a predictable amount of revenue in a predictable time. Texas Instruments and Fairchild Semiconductor invented the integrated circuit in parallel, not because they knew what the ROI would be, but because they believed that the opportunity would be tremendous. IBM's decision to move ahead with its PC despite the company's ROI requirements led to market leadership for nearly two decades. On the other hand, Xerox's Palo Alto Research Center invented laser printers, Ethernet and graphical user interfaces but ended up seeing its inventions commercialized by others.

Companies that hold every project to a strict Return on Investment calculation are likely to create only incremental improvements to existing products and processes--lots of "singles and doubles." Companies that are willing to ignore ROI in search of a greater goal are the ones that at least have a chance for true innovation--the "home runs" that can define, or redefine, an industry.

When "a better mousetrap" isn't enough

You've heard the saying "Build a better mousetrap, and the world will beat a path to your door." It's the mantra of many engineering-driven organizations. Unfortunately, it's not true. We all know examples of products that clearly were technologically inferior but that went on to great market success. In the U.S., one of the best examples was Sony's Betamax vs. Panasonic's and JVC's VHS. Most industry observers felt that Betamax was the better product--it certainly had better video quality. However, Betamax had two quality modes that allowed either one or two hours of recording on a single tape. VHS, on the other hand, had three modes that allowed one, two or three, and eventually two, four or six hours of recording. The picture looked better on Betamax, but consumers could purchase many fewer tapes with VHS, so it was seen as a much better value.

Sometimes, the most important innovations from technology companies have little or nothing to do with technology. In my opinion, Microsoft's two greatest innovations, the two things most responsible for its success, were software suites and per-machine pricing.

In the early days of the PC industry, consumers purchased applications one-at-a-time, based on their needs. If you wanted a word processor, WordPerfect and Wordstar were the preferred choices. The most popular spreadsheets by far were initially VisiCalc and then Lotus 1-2-3 and Borland's Quattro Pro. Microsoft had its own word processor, Word, and its own spreadsheet, first Multiplan and then Excel, but neither one was overtaking the market leaders. Then, Microsoft had the brilliant idea of bundling all of its office productivity applications together and selling them at about the same price as a copy of its competitors' single-purpose applications. The result was Microsoft Office.

Consumers immediately saw the value in Office. They might have preferred WordPerfect as a word processor or 1-2-3 as a spreadsheet, but for the same price, they could get a word processor, a spreadsheet and a presentation tool (PowerPoint.) Microsoft Office and its applications quickly dominated the market. Competitors tried to respond by acquiring other products to create their own suites, but Microsoft's market dominance was never challenged.

Microsoft's second innovation was per-machine licensing. Let's say that you were a large PC manufacturer, and you had a choice of a variety of operating systems--in particular, Microsoft's PC-DOS, Digital Research's DR-DOS and IBM's OS/2. Each one of the companies would sell you their operating systems at a price based on the total number of copies that you purchased. However, Microsoft came up with a unique new pricing model based not on the number of copies of PC-DOS that you shipped but rather, the number of computers that could run PC-DOS that you shipped. It cost much less per unit to license PC-DOS under this new model, but you had to buy a copy for every computer you built that could run PC-DOS.

This model almost immediately squeezed Microsoft's competitors out of the business of selling to computer manufacturers. PC-DOS was the "industry standard" and customers expected it. If you also wanted to offer DR-DOS, which many people thought was superior to PC-DOS, you had two choices:  Buy only the copies of PC-DOS that you needed, at a much higher price that you'd have to pass on to consumers, or buy two operating systems--PC-DOS for every machine, plus DR-DOS for some models. Very quickly, manufacturers decided that PC-DOS was good enough, and it wasn't worth raising prices or buying two copies of operating systems and throwing one away in order to offer a choice.

When Microsoft launched Windows, which was originally an add-on to DOS, it did the same thing: Manufacturers who wanted the lowest prices had to license DOS and Windows together for every machine that could run them. Competitive graphic environments such as GEM and Go didn't have a chance. This pricing model, more than anything else, built Microsoft's monopoly in desktop operating systems.

Have you ever wondered why Intel processors have been used by the vast majority of computer manufacturers for years, even when AMD had equivalent (or better) processors at lower prices? One big reason was that Intel was paying computer manufacturers under the table not to use AMD's processors (a fact that was recently admitted by Intel and Dell,) but Intel had another, above-board tool for getting buy-in. That Intel "Bum-bum-bum-bum" sound that you hear at the end of many PC commercials? Those commercials are paid for in large part by Intel. Through the use of co-op agreements and spiffs (sometimes called "sales promotion incentive funds), Intel reviews the commercials, and if they're approved will often pay 50% or more of the cost to air the ads. For the PC manufacturers, it's like doubling their advertising budgets. Intel won't approve payments for any ads that mention any computers using a competing processor, so it's a strong incentive to stick with Intel.

That brings us to a current example: How much is it worth to put the "Google" logo on your smartphones and have access to the Android Market? If you're Motorola or Samsung, it's apparently worth quite a lot. According to a lawsuit filed last week by Skyhook, a geopositioning technology company, Google withheld its approval for usage of the Google logo and access to the Android Market in order to force both companies to drop Skyhook in favor of Google's own positioning services. According to Skyhook, Google operates an "Android Compatibility Program," and products must be approved by this program in order to carry Google trademarks, license Google applications and gain access to the Android Market.

The Android Compatibility Program has two components: The Compatibility Test Suite, a software test suite that tests whether the submitted hardware and software are compatible with published Android specifications, and the Compliance Definition Document, which has additional requirements for what constitutes full compliance with Android specifications. According to Skyhook, the Compatibility Test Suite is an objective test that can be run by manufacturers and gives "go/no-go" answers, while the Compliance Definition Document is an amorphous, subjective document that can be freely interpreted by Google employees.

Skyhook claims that when Motorola submitted a phone that incorporated Skyhook's geolocation system, Google demanded that Skyhook share its geolocation information with Google in order to get approval. When Skyhook refused, Google then demanded that Motorola's phone run Skyhook's and Google's own geolocation systems simultaneously, which would have used far too much power and would have been impractical. Google additionally demanded that whenever the Skyhook system was in use, the phone's user had to be warned that their location data was going over a third-party network and might not be secure. After Skyhook refused to implement this final specification, Google demanded that Motorola remove the Skyhook system completely from its phone in order to get certification, and Motorola complied.

A second company, named "Company X" in Skyhook's lawsuit that is most likely Samsung, also adopted the Skyhook technology, and initially received shipping approval from Google. However, Motorola learned of the decision and requested that it be allowed to reinstate Skyhook's technology, at which point Google withdrew its approval to ship and reinstated it only after Company X removed Skyhook's technology from its phone.

As you can see, it's usually the exception when the better mousetrap wins, not the rule.

Well, the one on the right was on the left, and the one in the middle was on the right, and the one on the left was in the middle, and the guy in the rear used to work for IBM

Lots of job changes yesterday:

• Tony Fadell left Apple. Fadell was the first member of the iPod engineering team and most recently ran the iPod product line.
• According to CNET, Fadell is to be replaced by Mark Papermaster from IBM, but IBM is enforcing a non-competition clause in its employment contract with Papermaster, and is suing Apple to keep it from getting trade secrets related to IBM's Power chips and server products. What Power chips and servers have to do with portable media players is anyone's guess.
• George Kliavkoff, who's been NBC Universal's Chief Digital Officer for the last two years and oversaw NBC's participation in Hulu, the company's online Olympics activities and the growth of NBC.com and the company's other digital properties, is leaving the company to go do something else.