The software industry has gone through many revolutionary changes…and buzz words to match. After the tech bubble burst, software companies started to emphasize services – a more stable revenue stream that didn’t require pricey new product purchases from their customers. Enter the buzz word “Software Solutions”… fast forward to circa 2007… enter buzz word “Cloud Computing”…Here’s the chart:


If you rewind the above graph back to March 2008, could you predict the crossover? Is the ability to forecast such a shift useful? Basic predictive analytics can help you with the first question, you know the answer the to the second question – let’s help you with the first question.

The below graph shows how simple forecasts can be effective – it predicted the crossover to happen 8 weeks from the last data point, while the actual crossover happened in 7 weeks. The good news is that these basic forecast models can be implemented in simple spreadsheet programs. The honest news, is that if you select the wrong forecast model, your forecast may hurt you. The best practice is to have a conservative model and your data consistently updated – both of these tactics will keep big mistakes unlikely, while improving your decisions.

Simple forecast models for keyword trends