The Future of the Company

It’s fairly common to hear the CEO of a company that makes most of its money doing thing-X announce that the future of this company is thing-Y.

Why might they do this? I think it’s often just to sound visionary and bold. In rare cases it might be a clever plan to distract thing-X challengers by redefining the competitive agenda to include both things X and Y (Oracle excelled at this with many failed initiatives such as the NC).  There are certainly also cases where markets go bad and must be exited, such as when Intel needed to flee the commoditizing DRAM market.

But in the embrace-change-or-die culture of Silicon Valley, there’s also something fashionable about saying, “I know we are the leaders in thing-X. But thing-X is not strategic enough. So the future of this company is thing-Y.”

It takes guts to say it. TechCrunch and the VCs will get goosebumps.  There’s a certain burn-the-ships attitude that one can’t help but admire.

But is a good idea? I think, for the most part, no.

While it is sometimes absolutely necessary to lead companies through major transitions, I think the “future is Y” tactic is overused and often misapplied.
Sometimes burning the ships drives a high level of commitment to developing a fertile new world. Sometimes, however, it leaves a bunch of starving sailors on a barren rock.

My two favorite examples of this principle are Ingres and MarkLogic.

Let’s do Ingres first. Ingres was founded in 1980 about 2 years after Oracle and was one of the original 4 players in the relational database market. For a bevy of reasons more related to marketing and strategy than technology, Ingres lost the early relational database wars to Oracle. For example, by 1992, Ingres was struggling $240M division of the $400M struggling applications vendor ASK, while Oracle dominated the market and was about $1.2B. Somewhere around 1989 or so, it became clear to Ingres that they couldn’t win the core RDBMS market. In an epic display of bad decision-making, the company declared that the RDBMS market was commoditizing and decided to strategically focus on application development tools. I remember hearing top management say “the future of this company is application development tools.”

This was a terrible decision for several reasons:

• The RDBMS market was still in its infancy. Today, it’s a $15B (per year) market. Literally, hundreds of billions of RDBMSs were sold subsequent to Ingres’s decision that it was a poor market. It was in the top-two market opportunities for IT in the last century. (The other being PC operating systems.)

• The RDBMS market was not commoditizing. Ingres confused the growing dominance of a standard query language (SQL) with product commoditization. The products were all quite different. SQL would be extended and, in effect, re-made proprietary. The market wasn’t trending to pure competition and commoditization; it was headed towards oligopoly. Oracle drives 50%+ margins in RDBMS.

• The notion of carving a segment off the market was never considered. When someone loses the battle for dominance in market A, you don’t need to move to market B; you can segment the market instead. For example, Ingres was always very strong in query optimization. As Ingres was bailing out of RDBMS, a new multi-billion dollar segment was opening in databases specific to data warehousing, where that optimization technology would have been critical. Open source was another option was hiding in plain sight (the original University Ingres was always open-source; even before the term or concept was widely in use).

• Application development tools were a non-attractive market maybe 1/5th the then-current size with low barriers to entry, boatloads of competitors, strong downward pricing pressure (e.g., free runtimes), a powerful entry Visual Basic / Visual Studio and a bevy of other PC-based tools.

• The statement, and accompanying reallocation of resources, alienated all the database people who thought “why do I want to work at a database company not committed to databases?” The answer was you didn’t. Many of them ended up at Oracle and Sybase.

Tools were indeed the future of Ingres and that future ended up pretty bleak. In 1994, the whole struggling ASK/Ingres mess was sold to CA for $311M, a fraction of annual revenues.

Now, let’s look at MarkLogic. MarkLogic was founded in 2000, to create a hybrid DBMS / search engine based on the XML data model and the XQuery language. In 2003, Gartner wrote a note called XML DBMS: The Market That Never Was. Despite that, I joined the company as CEO in 2004 because, while I was aware that there was no horizontal momentum for the XML DBMS category, the company did have outstanding technology, strong investors, a great team, and a handful of good early customers.  I strongly believed it was a classic case of where a “bowling alley” strategy could be applied successfully.

(A bowling alley strategy is a systematic vertical market development strategy described in Inside the Tornado by Geoffrey Moore. The idea is simple: in the absence of strong category momentum, a vendor can be successful nevertheless by focusing on specific needs in specific industries and then bridge across markets by solving similar problems in new industries or new problems in the same industries.)

To me, it was clear cut and, God bless Geoffrey Moore, it worked. Of the dozen or so XML DBMS vendors in existence in 2004, the only one who succeeded in building a real business was MarkLogic. Some went out of business. Some repositioned into XML publishing applications. Some were sold for a pittance.  Do you remember any of these names:  Tamino, Ipedo, x-Hive, TigerLogic, Ixiasoft, Xyleme, eXist? It wasn’t exactly a hot category.

Despite the fact that 90%+ of the revenue was coming from the media and government verticals, several “important people” persisted in believing that the future of the company was enterprise. (Meaning, selling horizontally to F1000 IT.)  I never liked that because a horizontal enterprise assault had played no role in the company’s success to that point, and the data I saw suggested that wasn’t going to change in the future. To me, anyone paying close attention to the present might well conclude that if enterprise were the future, then that future might be pretty bleak.

Several things happen when leaders of thing-X companies declare the future of the company is thing-Y.

• You make counter-intuitive investment decisions. For example, if I told you that thing-X salespeople sold $2M/year and thing-Y salespeople sold $1M/year, you might expect that a company would consist of 100% thing-X salespeople. (For a given number of salespeople that maximizes revenue.) But once thing-Y is declared strategic, the company will seek to hire more of the lower productivity salespeople to support the strategy.

• Thing-X employees feel disenfranchised. I’d always felt that Oracle never forgot it was a database company. No matter the timeframe,  a top database engineer was a prestigious job. At Ingres, after about 1989, it wasn’t prestigious to be in the database group. Oracle never said tools or apps were the future of the company. Oracle, in effect, said: we are the leader in database and will leverage that to expand into tools and apps.

• Thing-X customers can potentially get rattled. I’m told that at the first MarkLogic User Conference after I left, that new management explained to an audience almost composed of media and government customers that the future of the company was enterprise.  It’s risky to treat your customers like your high school sweetheart on the day you went off to college.  (Hasta la vista.)

• Your best employees will want to move to thing-Y. By saying the future is thing-Y, you are announcing to your team that the past is thing-X. Your best and brightest will quickly get the message that to maximize their career opportunity, they should be working on thing-Y.

I firmly believe that there are situations where companies must leave an old market and enter a new one. Cognos’s exit from mid-range application development tools into BI is a great example. In 1996, you could safely say that the future of Cognos was BI. I’m sure that created many transition issues internally for them and, for the record, I believe that Cognos managed those issues extremely well. But given what was happening to VAX/VMS and MPE/XL Cognos had no choice. They needed to burn the ships and move to a new world.

But for, the most part, I believe that VCs, board members, and executives make one of two mistakes when they say the future of a thing-X company is thing-Y.

• They are hoping the future is thing-Y because for some reason they find that market more attractive than the market for thing-X. Often this may be a case of boredom with your own market (i.e., grass-is-greener syndrome) or a lack of creativity in developing it. Recall that Google was in part born by Yahoo declaring search non-strategic, and instead choosing to focus on portals and content.

• They are creating an artificial rhetorical “or” for dramatic purposes when the real message should be “and.” Follow the Oracle model instead:  we are going to lead in databases and, by hook or by crook, we are going to build an applications business because we believe that both are strategic to our future.

So the next time you hear anyone say the future of your company is Y, challenge them. Ask about X. Ask about Z. Ask about X and Y as opposed to X or Y. Then consider the real consequences of having your entire organization — and customer base — truly believe that the future is Y.

If you like that picture, stick with it. If you don’t, think a bit harder and change the plan.