Software module providers: winning at the first pass the post system*

In this post, I try to define a few rules to determine the probability of success for vendors of mobile software modules given the market conditions.


I would like to try to clarify an idea I have been discussing around about the strategic game that a software vendor can play. Let me explain. In the mobile industry, various vendors compete to offer software modules to manufacturers and ODMs, such as MMS clients, WAP browsers, vector graphics engines, network stacks, Java VMs, etc. I tend to think that in each category, the number of players who are eventually successful must be very small, probably 2 to 3.

As long as the market is not mature enough, all candidates compete. As soon as it matures, the market coalesces around two or three players, and that's the end of the game for the others. Once the top spots are taken, it's almost impossible to open new slots, because buyers are very conservative and tend to buy from the leader: success is self-reinforcing (see Geoffrey Moore's Inside the Tornado book). Why then isn't there only one player? The reason is probably that very often, all clients open the competition roughly at the same time: this industry is amazingly synchronized for that, and this makes it almost impossible for a single company to address globally and immediately. So my hypothesis is that there will always be at least two or three players in each category, but not more.

Now what does that mean to the vendor, the analyst and the investor? It means that it should be possible to define a few questions to analyze a vendor's potential for success. Let's take a crack at it and define the questions to ask...

Question 1- Has the proposed function already been identified as necessary on a platform? For instance, everybody knows what a SVG player could bring to a platform, but does that mean manufacturers see it as an opportunity? Or do they look at it as yet another annoying feature to implement? Similarly, a TTS (Text to speech) engine is nowhere talked about (as far as I know, I would love to be proved wrong...). Not that it is technically impossible (there are efficient engines available) but that it's just not in the map right now.

Question 2- Has the proposed function the ability of being implemented fairly autonomously, or does it have to be part of a greater module? Does it depend on another technology or module to be implemented? The idea here is to check any dependency that might hinder the deployment of the technology/function/product. For instance, a MP3 player is independent from other functions, but a MIDP game needs a Java VM. A WAP browser is also dependent on the platform, which allows some vendors to extend their offering, and therefore their stickiness: this is what Openwave has done with its browser by adding a stack layer that ties it to the platform; it becomes difficult to change the browser without important changes to the platform.

Question 3- Can the function easily be outsourced by the manufacturer? For instance, the graphic toolkit is too core to be outsourced. A tight integration to the hardware platform will bring too many benefits to be disposed of.

Question 4- How many suppliers are there for this function?

• 0: this never happens; if the manufacturer needs it, it's because it does exist somewhere. Manufacturers are now used to buy independent functions from third party vendors. This reduces the development time and the technical risks while minimizing the organisational impact. Not that they are not innovative, but they are not proactively looking at integrating new stuff because they, unlike many suppliers, know what it costs.
• 1: in this case, manufacturers will evaluate the supplier, but will be looking for alternatives; this might mean that there is no urgency yet for the function, and that manufacturers can wait and extend the "let me think about it" period (the valley of death for the suppliers). For the suppliers, however, it means that the race is open and can heat up anytime. If a vendor is not in the game by that time, it might be too late unless it has a really good go-to-market strategy.
• >1: in this case, the manufacturers will choose from the existing offer and tend not to look outside this group for alternatives. The game is already over for a new entrant, and even if some sales can be achieved, it will certainly only be on the periphery of the market and hence be not very profitable and growth generating.

Question 4 can be considered by looking at the maturity of the vendors:

1- Vendors whose product has already been integrated by a at least one important manufacturer - and even better, deployed - clearly are in a different category. Only they can have the credibility to race ahead for the leadership. This also has the effect of ackowledging the importance of the function for the whole market.

2- If no manufacturer has deployed such a function, the market is still open. Until a manufacturer has implemented it, the function does not seem so critical, and the sales cycle is very long. When it is, all manufacturers consider it. That is when the market accelerates and the race really starts, the sales cycle shortens dramatically. Vendors who are not in an advanced stage in terms of integration and deployment are in trouble.

The implication of all this seems to be that there is a "first deployer" advantage. The vendor who deploys first can win because it will be seen as the future leader. Alternative offers will be evaluated, but they will face an uphill battle, each sales becoming more expensive.

So the "rule of thumb" might be: on any give market, if two vendors already have deployed, the game is over for the other candidates, unless some disruption re-opens the game.

In a next post, I will try to apply this framework to a few sectors...