There are (at least) two major short-comings to today’s forecasting approaches. First, forecasters are using increasingly narrow definitions so they aren’t seeing things in the broader context. Secondly (and related), forecasters largely fail to look beyond what is seen – many of today’s forecasts are just projections of the current state of affairs. Taken together, today’s forecasters are missing the broader landscape of innovation that is taking place.
It is important to remember that adoption of innovation – whether it is consumer adoption or enterprise adoption – follows s-shaped power law curves. Growth isn’t linear. Adoption typically starts very slowly and remains low for an extended period (ie several years). After several years of low to no adoption, grow accelerates at an accelerating rate (positive 2nd derivative) for an extended period. Eventually adoption slows – the 2nd derivative turns negative followed by the 1st derivative turning negative.
This is where today’s analysts tend to stop. The drive towards specialization forces forecasters to focus on very narrow definitions. Analysts tend to take a micro approach to forecasts and focus only on single s-shaped adoption curves. They look at things in isolation. They look at markets in two states – growing or shrinking.
In reality, developments materialize over long stretches of time and typically overlap other trends. A single s-shape adoption curve will beget dozens or even hundreds of other s-shape adoption curves. In this way, innovation is much larger in aggregate than forecasters convey with their narrowly-defined forecasts. We think in narrow terms. In reality innovation follows k-waves as opposed to standard normal curves. More, k-waves overlap. Because these developments occur over long periods of time most companies are typically able to adjust to the evolutionary process of adoption. They are able to move to areas with positive 2nd derivatives, while abandoning areas with negative 1st derivatives. This transition doesn’t always happen smoothly, but it does happen naturally. But the desired immediate applicability of forecasts fails to accurately account for this evolutionary process.
