Curious which promotions lead to sales? We can help.
“I know only half of my advertising works. The problem is, I don’t know which half.” – John Wanamaker
NOTE: Relative conversion score has been upgraded! Please see our blog post about how RCS has morphed to TCS as of November 23, 2015.
One of the greatest challenges in online marketing is understanding which of your efforts are actually leading to sales. Closing this “attribution loop” is the focus of many marketers. Marketers whose sole purpose is to answer questions like What social media platforms drove sales? What products are people actually buying? Which marketing channel has been most cost-effective?
These questions are even harder to answer when you’re marketing items through a third party service, such as iTunes. Thankfully, the iTunes / App Store Affiliate Program does provide some insights, but not without significant challenges and limitations.
As the leading link management platform for the iTunes ecosystem, attributing clicks to sales has been a riddle GeoRiot has been working to crack from day one. Since then, we’ve been quietly experimenting, creating numerous iterations and going back to the drawing board countless times to figure out the best way to close the loop. Today, we are ecstatic to be pushing live our Relative Conversion Score to finally help our clients answer the questions that mean the most to their bottom line.
What is the Relative Conversion Score?
The Relative Conversion Score (“RCS” for short) is a value that we’ve added to all of our iTunes reporting in the GeoRiot dashboard to help you find which aspects of your marketing work great and which ones aren’t doing so well.
To get the most value from these scores it is best to use them as a comparison between the different sources of data in a report. They should not be confused with a “conversion rate” and provide little relevance as a standalone value.
Comparing Relative Conversion Scores.
For example, in the screenshot above we could guess that the band’s website was about 50% more effective than the label’s website in promoting a specific product. Further, we could quickly extrapolate that Twitter was about three times more effective than Facebook while both significantly outperformed YouTube.
“Conversion” reporting for the iTunes / App Store Affiliate Program first came to light when iTunes switched to use PHG exclusively. The “Conversion Rate” metric was introduced into the PHG dashboard and was calculated simply by taking the number of items sold divided by the number of clicks. While (it was) a great start it was a bit of a misnomer as the items sold could be anything that is bought during the 24-hour affiliate window and is not limited to the specific item you are promoting. For example, if you link to an album and had one click that resulted in the sale of three free apps your conversion rate would still be 300%.
Calculating the RCS is substantially more complicated but results in a score that is significantly more accurate. Here’s why:
- We use three different data sets to calculate a score per click (not per link or per metric). This allows us to analyze every click, link, and sale (instead of in aggregate) to avoid “false positives” that would inflate values. This also allows us to assign the RCS value to each of the different attributes we report on (country, referrer, media type, device, etc.).
- We introduced a time-based decay algorithm to reward purchases that were made shortly after the click. For example, if a purchase is made a minute after a click, it’s more likely it’s the item you’re promoting while an item that is purchased 23 hours after a click is less likely.
- We’ve built in rules to handle multiple purchases of the same media type and price as you can’t buy the same product twice (after you purchase once, it’s yours forever).
- We also developed detailed rules to account for pre-orders, apps with in-app purchases, products that had “upsell” opportunities (click a link to a song, but buy the album instead), and items available in multiple formats (Ex. HD, SD or rental) to provide better insight.
It’s important to understand that there are inherent limitations from using this approach. The affiliate program does allow us to get sales data from iTunes, but there are still three significant challenges:
- The data coming from the affiliate program is limited to the “Category” level. Specific items are not reported, only the media type of the item.
- The affiliate window is a maximum of 24 hours and can be even shorter if the user clicks another affiliate link before purchasing.
- The iTunes affiliate data needed to find the conversion score is often delayed 3-5 days but sometimes multiple weeks.
As with our matching algorithms, we’ll continue to tweak and tune our RCS algorithm and with your help, we can continue to make the metric even better.
Where are my Relative Conversion Scores?
If you don’t already see your relative conversion scores live on your dashboard it’s easy to set up this new reporting. Simply:
Now that you know all about our new metric, go see what it can reveal about your marketing! After you’ve gotten familiar with the RCS, we’d love to hear your thoughts. We are always looking for ways to improve our platform and learn the ways in which you derive the maximum value from our service.
Thank you, as always, for your support.