Never believe a single data point when forecasting

We often receive questions such as “researcher X’s forecast is much higher/lower.”  I always take these remarks seriously – there are many elements an analyst has to take into account in forecasting and  I’ve learned over time that the content of these conversations drives the value of the forecast (for both parties).

Recently I had a discussion with a client who was skeptical of our growth projection for digital music subscriptions in our recent Music forecast.  We discussed that a key input to our estimate are discussions with companies providing these services and understanding their growth outlook.  But these discussions are expectedly biased, that’s why we routinely challenge providers to defend their expectations but we also look for corroborating data to indicate direction and scale of change. Similarly, if survey respondents happen to love a product idea, it has potential but there’s no certainty that consumers will eventually spend money on that product.  What we look for is a solid pattern of evidence supporting a growth hypothesis along with a paucity of evidence supporting the alternative (decline).

As part of the forecast process, we as analysts debate various hypotheses and I shared the content of those discussions with this client, which helped him understand that our growth expectation for digital subscription is supported by evidence beyond the confidence of subscription providers.

In this vein, we found it complimentary to have another part of our Music forecast described in this way by an industry publication, that stated “Given the realities of the marketplace (people still like CDs, labels know they must accommodate concerned retailers), this is a more believable forecast of physical revenues.”  Over time, our process has been proven to deliver accuracy. But for our clients our willingness to share the process and discuss the decisions made are equally important, and have made the product even more valuable.