This is the summary of an article by Nicholas Lincoln, which argues that not all product’s sales revenues can be predicted by keyword search volume, citing various factors that come into play. It also states that there is no correlation between ad spending and sales revenue for the Apple iPod and the iPhone. You can get the pdf of the original behavioral targeting article here: The Relationship Between Internet Marketing, Search Volume and Product Sales.
Related Studies: Box-office Forecasts Using Twitter
A study in 2010, sponsored by Hewlett Packard, shows that Twitter sentiment and rate for movies can be used to outperform market-based predictors for predicting box-office outcomes. This shows that social networking sites can be used to make real word results predictions.
In this study, the measure of popularity used is not Twitter, but Google search. It studies search volume that spans six years (January 2004 to December 2010).
Two key resources were used in obtaining the data used in this study: Google Insights for Search and Apple’s Form 10-K reports. Both Google and Apple were used for their acceptable global search volume representation, and popularity and brand awareness, respectively.
The top twenty keywords containg iPhone and iPod were obtained from Google Adwords. Search volume data on a weekly basis was collected for the keywords. Since Google Insights for Search is scaled, exact numbers were not used.
One hundred is given to the keyword with the highest search volume for a given week, and the rest go in proportion. Finally, from the Form 10-K reports, quarterly sales revenues were obtained for the iPhone and the iPod.
Setting up the Hypothesis Test
There are two parts to the experiment. First, determining how advertising affects internet popularity, and second, how internet popularity affects sales revenue. A correlation discovered for both parts implies that advertising affects sales revenue. A combined model should then include advertising and internet popularity.
The model developed in this study is described as a linear finite distributed lag model created with the present and future effects of internet popularity and advertising in mind.
Results and Discussion
Amazingly, this study shows that advertising did not affect sales revenues for both the Apple iPod or the iPhone. This may be due to the fact that Apple’s marketing strategy is quite unique; the company relies on the all-consuming loyalty that consumers have toward the Apple brand, by letting them try the devices and spread their existence via word-of-mouth.
Apple only increased advertising if sales revenue lowered, but sales continued to be high enough that it didn’t become a good factor to determining sales revenue.
Another surprising result showed that internet popularity had no effect on the sales revenue of the iPod. This can be explained by the fact that the iPod was already a well known product during 2004 and the price remained relatively the same throughout the six-year period. Google search towards the iPod was unnecessary because people usually search for these devices if they want to know more about them.
In contrast, there was a strong influence by internet popularity towards the iPhone during the year 2010. There was a peak of interest for the iPhone at this time because of the announcement of the new service provider: Verizon Wireless.
As such, the results of this study were inconclusive, as one product was not affected by internet popularity, while the other was. It is recommended that this study be done on a larger scale, with more products, to determine if internet popularity can be used to predict sales revenue, which can then be taken advantage of by advertisers.