Rely on Third-Parties or Do-It-Yourself?

Further to a recent post (“Activity and Inactivity“)…

In the last few days, Twitter has made aggressive moves in an attempt to reign-in the set of third-party Twitter clients. They have changed their terms-of-service to discourage / forbid tools that replace the official Twitter software. As per Twitter’s Ryan Sarver:

…developers ask us if they should build client apps that mimic or reproduce the mainstream Twitter consumer client experience.  The answer is no.

There has been uproar amongst developers, on the grounds that Twitter has been successful in significant part because of innovations amongst third-party software providers. Even today, 42% of all Tweets come from third-party applications; now Twitter appears ready to cut the ground out from under the developers of those clients.

On the other hand, you can see why Twitter feels it has no choice. Twitter does not own the social content people post to it – it belongs to the posters, and they could easily put it somewhere else (FaceBook, LinkedIn, StumbleUpon, etc.). Nor does Twitter own the content to which Tweets typically link. Twitter’s assets with a half-life of more than a day are:

1. The graph of interactions between users, and perhaps between users and content
2. The user experience that people are familiar with
3. The size of their user-base and the service’s visibility with news media

By letting third-party clients proliferate, Twitter lost control of #2 – user experience – and made it much harder to monetize #3 – eyeballs.

At the same time, that proliferation did create all kinds of innovation and allowed the service to scale while lessening the complexity and risks and costs of scaling for Twitter itself.

Sometimes, that is a trade-off that is worth making. It is important to remember, though, that what looks like a great idea early-on – gaining leverage from third-parties – can seem a lot less smart if it makes it harder/impossible to scale a business three years later. Sometimes, indeed, raising more venture money and managing the organizational complexity of more employees, even early on, is in fact the best way to grow. My guess is that that would have been the case for Twitter, given how “hot” they were (and are) and their consequent access to capital on very favorable terms.

As discussed in the last post, though, the “Let’s build it ourselves” approach would have required a lot more innovation and experimentation by Twitter, and perhaps they didn’t feel equal to that at the time. Even then, though, they could have thought-through the business-model implications of an open-API – for instance by taking a cut of all advertising revenue associated with the Twitter API.

What are the lessons for the various startups that will be damaged by this change in attitude by Twitter?

• Building a startup predicated on the ecosystem of a single-company can be dangerous, especially if the company has yet to fully establish it’s own business model – in case the company decides its business model looks a lot like your business.
• If you are heavily dependent on one or two major players, try and be seen as a friend not an enemy. For instance, as Twitter pushes more advertising, third-party clients could look for ways to voluntarily support Twitter’s own advertising.

In future, clients that are “a better mousetrap” than the Twitter client are probably unattractive startup opportunities. However, clients that might re-imagine social media, and/or spread across multiple forms of social media could still be very appealing. Twitter’s record to-date does not suggest that they will come up with the  last word in social-media user-experience innovation.