Testing for eCommerce | Part 2 – Quantifying Opportunities

Prioritizing Your Plan Of Action

In part one of this series, we covered strategies and techniques to identify areas of your site that are prime for testing and optimization. In this post, we’re taking it a step further to quantify these opportunities. This allows you to estimate the return gained from improving specific metrics in the relevant progression points that are affected.


Tying in Revenue

More than likely, as an analyst, at the end of the day you’re going to have to build a strong case for testing to present to a certain group of stakeholders. It may not be enough to talk about improvements to the UX, or changes to branding; at the end of the day what will capture the attention of interested parties is the bottom line. This means evaluating the areas of opportunity you have previously identified, and tying in revenue. Establish the importance of the micro-transaction targeted for testing, either by reporting the page value within Google Analytics or better yet, leveraging Google Analytics advanced segments to calculate conversion metrics for the specific segment you’re targeting.


Creating Improvement Models

Now that you have established a baseline for your areas of opportunity, the next step is to project potential return by introducing improvement models. This step can be critical. It may not mean much to your stakeholders to tell them you think that by redesigning the product page, you could potentially see a 11% lift to progression rate, but if you prove out in a simple business model that an 11% lift to this single micro-conversion translates to a potential $100K increase in monthly revenue sitewide, now you’ll have their attention.

There are many ways to go about this, but the simplest and most straightforward is to take existing metrics over a period of time, and apply your estimated lift to only the one single metric you’ll be targeting (progression rate in this example). From there extrapolate this change, again using existing metrics, to create an estimated return of investment. The one tricky part of this is determining just how much lift your changes will create. While this will come with experience, and knowledge gained from seeing how your user-based respond to your tests, it’s good practice to introduce low, mid, and high-effect scenarios. This provides a range of projected return, so you don’t find yourself quoting something that you can’t deliver on if the test goes south.



Prioritizing Based on Impact

More than likely, by now you have an entire list of areas of your site you’d like to test, and business cases for each. Obviously the next step is to prioritize this list so you can get started. In most cases this goes beyond simply ranking them in order of projected revenue increase. There’s a host of other factors to consider. Each business is different, but here are some things to think about:

  • Projected revenue increase
  • Estimated development time required

  • Estimated test duration

  • Difficulty to achieve desired lift

After examining these factors, you may find your test priority moving around a bit. In some cases, smaller tests which are projected smaller revenue increases may move higher up on your list because they require less effort on your part to execute. Sometimes it pays to go for the low hanging fruit.