How Long Will Your Business Last Series: Part 1
Can you predict the lifespan and effectiveness of a company just by looking at publicly available analytics? Surely not, right? Otherwise everyone would be doing it and trying to make millions in the process. Maybe. I’m not so sure.
I work as Data Science Lead at productOps, a small bespoke consulting firm focused on partnering deeply with clients to ensure their success, even if it means asking tough questions to make it happen. Part of delivering that service is conducting research on how companies and markets actually work so that we can understand our clients and their businesses from the ground up.
For a while now, I’ve been thinking of client organizations, and companies in general, in terms of a developmental arc they often follow. Generally, there is an accelerated, occasionally exponential rise, a high point, and a slower decline curve with a long, drawn-out tail.
I long assumed that ‘the arc’ would remain a simple model I scribbled on white-boards while trying to explain organizational change. However, a recent, simple discovery I made radically changed my thinking on this topic. How close was my intuitive picture of company development to measured reality, I wondered? While doing some research for a client assessing the current global state of the cloud computing industry, I found myself on Google Trends. Trends is a fantastic tool and a goldmine of comparative social data. I couldn’t resist taking a look, so picked a company whose rise and decline would tidily fall within the Google Trends data collection range and brought up its data.
On the face of it, this curve seemed a perfect match to the mental model I outlined above. But what exactly are we seeing here and why should we consider Google Trends data as significant?
I’d propose that Trends data serves as a reliable proxy for the social capital of an organization. Furthermore, if we pick a good proxy, it’s potentially a more robust indicator of organizational health than something like quarterly reports or stock price. This is because crowd-sourced social proxies like long-term attentional weight are very hard to game and don’t usually serve as the platform for financial speculation or industry signalling.
To check whether the pattern was a fluke or something genuinely interesting, I broadened my search to similar organizations. Here are a few of the arcs that I found.
Tech companies seem to provide a remarkably clear signal because their lifespans are often short and their social impacts frequently enormous. Not all of them perfectly fit this profile, of course. However, the outliers are also fascinating, precisely because of what they’ve achieved.
What about organizations that are too long-lived for Google’s data-window, you may ask? The answer is that the curves appear to be extended versions of exactly the same pattern. In these cases, we only get to see part of the elephant, but ‘the arc’ seems intact.
At this point, it only seems logical to wonder about the fates of longer-lived social institutions. What other kinds of indirect markers of economic success might there be? How about organization size, for instance, as measured by the scale of the engaged population? This kind of data isn’t usually accessible for companies. However, it is often readily available for company towns. So I took a look at the historic populations of Pittsburgh and Detroit.
This suggests, to me that what we’re looking at here is not simply a Trends artifact but a more general pattern applicable to a broad swath of human institutions and potentially other social constructs besides. Commonalities in related data domains often offer clues as to a common mechanisms that we can use to build predictive models so I decided to cast the net as wide as possible to see where else similar patterns showed up. As soon as I started looking, I found the pattern everywhere from the depletion of oil reserves to the sizes of bacterial colonies.
In fact, this pattern is so ubiquitous, I feel certain that it must have studied in several different contexts. But what’s not clear to me is whether the interdisciplinary implications have been fully understood. Even if the underlying drivers for human social systems are different from that of bacteria, it doesn’t appear to matter. Furthermore, there seems little reason to believe that different mechanisms are actually at work. At some abstract level, system constraints related to food/revenue sources and optimized growth may be identical.
Fortunately, modeling data on this scale is far easier than analyzing artifacts in the terabytes of information sometimes gathered by our clients. In fact, a very simple system appears to suffice to match the data remarkably tightly. This model doesn’t just give us a way of understanding organizational growth but also a tool for predicting organizational lifespan. Organizations that understand that mechanism can avoid decline so long as they’re prepared to look hard at the data and ask the right tough questions.
In Part 2 of this series, I’ll show you how that model works, explain how organisms in nature beat the arc, and show you how companies can potentially use that knowledge to lock in lasting growth.