9 years, 4 months ago

Towards a VPEC-T analysis of Google

#entarch Enterprise architects need to understand values and policies. VPEC-T is an approach that is particularly useful for situations where there are multiple conflicting values and policies, or multiple interpretations of What-Is-Going-On.

In this post, I want to look at Google. Can we infer its values and policies from its observed behaviour (over time).

We may start by asking what events we think Google is paying attention to. Here are some of the events that are available to Google.

1. You search for “XYZ”
2. You skip over the first few items, and click on the third item on the second page.
3. You look at a webpage and then come back to continue your search.
4  You rephrase and clarify your enquiry.

Google is pretty coy on its exact use of these events, but most Google-watchers assume that these events have an influence on its search algorithms and/or its advertising algorithms. In other words, we may presume that Google is generating valuable content from these events.

Google has indulged in a wide range of initiatives over the years, many of which have no obvious line to revenue. But all of them have the potential to generate vast amounts of rich content – much of it related to the observed behaviour of internet users. On this interpretation of Google’s strategy, initiatives are dropped, not because they fail to generate revenue but because they fail to generate enough of the desired kind of content. Google is betting its future on building and maintaining this content through powerful positive feedback.

Google’s strategy is therefore surprisingly traditional – it involves capturing some territory and defending it against its competitors. Here’s an example – Google provides the Android platform to mobile device manufacturers. When Motorola wanted to use Skyhook’s voice recognition instead of Google’s, Google forced it to fall into line. Daniel Soar argues that this was not because Google executives feared losing revenue but because they feared losing access to an important source of content. As Soar puts it, “Google faced the unfamiliar problem of the negative feedback loop: the fewer people that used its product, the less information it would have and the worse the product would get.” (Google has since bought out Motorola Mobility, which presumably resolves some of the trust issues as well.)

Daniel Soar, You can’t get away from Google, London Review of Books, Vol 33 No 19, 6 October 2011

Can we understand Google’s phenomenal collection and use of data as an example of organizational intelligence? Google is certainly seeking to differentiate each Internet user’s experience, as well as integrate across multiple domains (web browsing, email, blogging, voice, video, satnav, and so on). Google already has an army of brilliant engineers as well as an alarmingly large carbon footprint. There is lots of evidence of Google’s integrating these resources into one of the most innovative sociotechnical systems on the planet.

(By the way, when I asked Google itself about its carbon footprint, it recommended I look at a recent story in the Guardian (8 September 2011). I can see that Google has been asked this question many times before, because it pops up so quickly as an expected search term. But why should I trust Google’s recommendation, and how can I ever discover what newspapers would be recommended to a browser with a different browsing history to mine?)

But a lot of this learning looks suspiciously like first-order learning. So the content gets better, based on better capture of events, but to what extent is there any systematic evolution of policies or questioning of values? There may well be some second-order or third-order learning, but it’s not easy to see from the outside. There is also an important question about the relationship between Google’s own ability to learn from its accumulated content, and Google’s ability or willingness to provide a rich platform for learning by others in its ecosystem – in other words, a broader notion of collective intelligence.

I wonder if there are any lessons for other organizations? Sometimes firms like Amazon, Apple, Facebook and Google (Eric Schmidt’s Gang of Four) seem pretty far removed from most other organizations, but their platform strategies and operating patterns will surely become increasingly relevant in other sectors. A traditional retailer may now collect and analyse a much larger quantity of data about its customers’ behaviour than ever before, even if this is still several orders of magnitude less than what Google does. A traditional telecoms or media company may now see itself as a platform business in a multisided market. Therefore instead of seeing Eric Schmidt’s Gang of Four as impossibly remote and mysterious organizations, populated by unbelievably talented and creative engineers, we should start to think of them as harbingers of the enterprise of the future.

See also my post Google as a Platform (not)

 book now  Business Architecture Bootcamp (November 22-23, 2011)
 book now  Workshop: Organizational Intelligence (November 24th, 2011)