Tony Bain

Building products & teams that leverage data, analytics, AI & automation to do amazing things

The Chicken or the DB

August 06, 2008 | Tony Bain

The database platform market is a No Content: 18.8b market that is largely dominated by 3 players, IBM, Oracle and Microsoft. Every year though we see new products come onto the market that claim to be the next big thing or offer significant benefits over the traditional big 3 RDBMS’s. Often that may be the case, but the problem facing every entrant is that the big three’s dominance is due to historical timing and leveraging opportunities that existed which no longer exist.

IBM has its place in the big 3 due to a 30 year history in the relational market, in fact as the original developer of the relational theory and the SQL language. However IBM has largely focused on its mainframe and mini computers over the last 30 years and its database revenue is largely tied to continued revenue streams from these core systems.

Oracle again has a 30 year history, they really developed and pushed along the advancements of the relational database platform and so quite rightly are the kings of the database revenue castle. And they build their customer base from a top down approach, focusing on the big end of town first filtering its products down from large mission critical core database servers, though to smaller and mid range requirements.

Microsoft on the other hand recognized an opportunity for entering the market by taking over the small end of the scale and working itself up. In the mid 1990’s, SQL Server was launched and the focus was on workgroup, mid range requirements – essentially a step up from their Access product. Due to Microsoft’s dominance at this level, the PC level, SQL Server rapidly expanded and as people gained confidence and expertise in the product, SQL Server slowly worked its way up into the enterprise over the last 14 years. However the other card that Microsoft carried was the high level of third party application developer engagement and support. Making it easy to build new SQL Server database applications, and the lack of a “cheap” easy to use mid range database platform alternative meant that third party developers in their droves started building their mid range apps to use SQL Server instead of (or as well as) Access or Oracle.

So the situation today is that IBM, Oracle and Microsoft have all come to the same point with Oracle in the middle and are competing head to head to make incremental gains in their respective market shares. Oracle and IBM are competing on the non Windows platforms, Oracle and Microsoft are competing on the Windows platform. The products themselves are largely equivalent, so the competitive advantage is no longer about core features, corporations are making decisions on what is their preferred platform based on cost of ownership, in house skills, and customer/vendor relationship quality.

But why is it so hard for another vendor to enter this market with something that is better, cheaper, faster etc? Well the database market has a chicken or the egg situation surrounding the applications that use the databases themselves. If you look at database applications that are focused on use in medium sized businesses and above, you can be 99% sure that they will support either Oracle or SQL Server or more likely, both. For database applications that only run on Windows, almost all support SQL Server, many support SQL Server and Oracle a subset just support Oracle. Even Oracles own applications (that it has acquired) for example, like JDE and Peoplesoft support using Oracle and SQL Server databases. The reason is that the third party application vendor doesn’t want to lose sales because their product doesn’t support the chosen database platforms of their potential customer. They know if they support SQL Server and Oracle they will be compatible with almost every customer opportunity they come across.

You may then ask why companies are “choosing” a database platform, shouldn’t they just implement whatever is the best (or cheapest) for the application requirement? Well to reduce TCO around the database platform companies have found that the only way you can really do this is to have a high level of standardization on one platform or another as this gets them better license deals and allows them to keep a single core skill-sets in house to manage those systems. Note, most companies don’t fully exclude any other database platforms completely, instead when performing a database application selection process, applications which don’t work with their “preferred platform(s)” are marked down, so in theory they could still win assuming they have significant other advantages over competitive products (in reality this is rare).

For a database application vendor to add support for multiple database platforms is a costly process. There is development effort, testing effort, ongoing testing of database patches etc and if supporting Oracle and SQL Server covers almost all potential sales, how do you justify such high additional costs for such a low return that will be achieved by support other database platforms?
So the chicken and the egg? As I mentioned the situation now is that few database application vendors are going to build a product for a database platform that a large number of potential customers aren’t using. The customer on the other hand isn’t going to choose to implement a database platform that most of their database applications don’t support.
Ok, so as a database platform vendor how do you enter this market then? Well there are a few ways.

  1. You launch a product that is so compelling that customers start demanding their application vendors support it. Note, creating an equivalent product and making it “free” to date has failed to create this level of demand.
  2. You launch a product that is so compelling that developers build their applications for it and then force their customers to install the platform.
  3. You predict a future trend and get started in that future trend early before any of the main three do and create a level of dominance. Note, object databases as a future trend has a long trail of wreckage behind it. Clouds are interesting and BigTable & SimpleDB have gained some early dominance here.
  4. You forget the corporate bread and butter database market for the time being and pick off a niche or two and focus on that. Clouds (currently a niche but may be the next big thing), In Memory DB’s, data warehouse appliances, rotated tables. All nichey nichey, but allows a new vendor to gain some sort of a foot hold. But it will have to be a very big niche if your foothold is going to leverage you into the corporate database market, and it will have to be a niche that you will have to maintain dominance in for 10 years or more (based on past history).

So #4 seems like the easiest way to enter the database market and make some money, but of course it can’t be that easy. As I mentioned, the big 3 are kind of in a stalemate situation in terms of gaining market share over each other so they are looking at other areas in which to gain ground, which of course includes this niche market (and there have been a lot of niche acquisitions by Microsoft and Oracle of late)!