In the first blog post of this year, I talked about some of the challenges paid search managers face in terms of managing tests in their accounts. The issues, as that post discussed at length, include:
- There are a lot of interacting variables – it’s a complex system
- Critical data is hidden from us – we can’t see what’s happened.
- Our tools don’t facilitate testing – which is just crazy.
And the fact that we face two statistics problems.
- The math we need isn’t trivial
- The data we usually have is sparse and dirty
These are issues we think that good paid search software should address, and in the April 2011 release of ClickEquations we’ve begun with the introduction of our Text Ad Zoom feature.
Text Ad Zoom
Text Ad Zoom makes it easy to see an accurate analysis of the statistical confidence you can have in the CTR or conversion rate metrics reported for your ads. For years paid search managers have had to ‘eyeball’ or ‘estimate’ the statistical signficance of these numbers. Some have relied on ‘rules of thumb’ like ‘wait for at least 100 impressions’ or ‘wait for at least 10 clicks’ but most of the time we just take the numbers at their face value and if one ad has what seems like a dramatically lower CTR or Conversion rate than it’s peers, we pause it or rewrite it. With Text Ad Zoom the guesswork has been removed. By simply double-clicking on any text ad in ClickEquations Manager you’ll be presented with the current metrics for all ads in the ad group – using the data from the currently selected date range. The best performing ad is automatically highlighted at the top of the Text Ad Zoom dialog box, with all of the current alternative ads in the ad group listed below.
Click To See Full Size
Next to the CTR and Conversion Rate metrics is a simple rating for whether or not the reported metrics are significant. The ratings are Yes, No, and Maybe.
- When the significance rating is YES, the results are statistically significant. A ‘Yes’ rating means >95% confidence. In this case you can trust the numbers and take action on them. If the ads keep running the future results are likely to be consistent with the past results..
- When the significance rating is NO, the results are not statistically significant. A ‘No’ rating means confidence is <90%. In this case you should not make decisions on this ad based on the current data – there has not yet been enough data collected to trust that they reflect an accurate representation of what will happen in the future..
- When the significance rating is MAYBE, the confidence is between 90% and 95%, and you can probably trust the numbers and take action on them (it is likely that the current trends will continue) but if you want to be more certain then you should let the ads run to gather additional data.
More To Come
A built-in statistical significance test is just one of the ways we can make it easier to test ad copy and better manage your campaigns. Watch for more to come in future releases of ClickEquations. BTW, Text Ad Zoom is just one of the new features we added to the April 2011 Release of ClickEquations. We’ll look at other cool new features in upcoming posts.