Struggling With Data Discrepancy? Here’s How To (Start To) Fix It.

image

Measuring performance data through multiple platforms is an inevitable reality for most direct response driven organizations. There are plenty of reasons why having multiple analytics platforms is necessary, such as:

  • Keeping the performance numbers honest. If you are running a display campaign using an Ad Network, you would want to track it via an Ad Tracking platform as well. Similarly, if you are testing out a new Analytics solution, it makes sense to have another one running side by side just to compare the numbers and spot anything ridiculously out of sync.
  • One platform cannot provide all the metrics. If you are running an ad campaign directly on the native platform, there are certain metrics, like view-through conversions, which might not be measured by your Analytics platform. The discrepancy here can arise when you measure a common metric provided by both the platforms, like clicks.

Dependence on multiple platforms almost always leads to some amount of data discrepancy. Here are some common reasons why they occur, and how to fix them.

  1. Time zones don’t match, aka Most Common. Ensure that all your platforms are setup in the same time zone. Failing to do so will lead them to report metrics captured over different hours/days. This can cause huge discrepancies if one platform is set in PST and the other on the other side of the planet in SGT. Also, it’s critical to note that some platforms don’t allow you to change your time zones in the same account post-setup.
  2. The tags are not placed close together. How many times have you accidentally clicked on an ad, only to press escape or the ‘back’ key to stop the page from loading? Sounds familiar doesn’t it? What you have essentially done is allowing only a partial HTML code of the website to load. This means that some of your tracking tags have fired, and some haven’t.

    If you return to the website within the cookie window (a date range within which any conversion is attributed to the click) and make a purchase, that conversion will be attributed to the tracking platform whose tags loaded in the past. The result, one platform gets the conversion, while the other does not.

    The best way to avoid this is to place the tags inside a tag manager like Adobe’s DTM or Google’s Tag Manager. If your organization does not allow the use of tag managers, ensure that the tags are placed right after each other. If possible, place them right before the </head> tag.

    data html discrepancy
    While this will not eliminate the discrepancy completely, it will definitely narrow down the difference.

  3. The cookie windows are not the same. A cookie look-back window tells the tracking platform about how long it should consider the validity of a click or impression. For example, if you set the cookie window on AdWords to 30 days, and if someone clicks on an ad today, that particular ad will be attributed all the conversions happening over the next 30 days for that user.

    data html discrepancy
    Clearly, setting different look-back periods on different platforms can potentially cause a massive discrepancy.

  4. The Attribution Models don’t match. Attribution decides how much weightage to allocate to an event (click, impressions, engagement) leading up to a conversion on the website. While a lot of organizations still rely on last click attribution, there are many sophisticated organizations which look at other models like even distribution. This ensures that contributing channels like display get their share of the conversion as well.

    Here is an example of how attribution would work in the two scenarios.

    User Journey: Impression (AdWords) -> Click (Facebook) -> Click (LinkedIn) -> Click (Email) -> Conversion of $100

    Last click attribution = $100 to Email
    Even distribution      = ($100/4) = $25 to all the channels involved.

    If you don’t have the same attribution allocated to all your tracking platforms, you can see clearly how the conversion numbers would fluctuate wildly.

  5. Post-impression conversions are not universally detected. A lot of display and social platforms report not just conversions after an ad click, but also after an ad impression. These conversions are not always separated out during reporting and due diligence will ensure that this does not happen. Most Analytics solutions cannot track post-impression conversions unless they are manually fed into the solution.
  6. Currencies don’t match. Ensure that you are feeding data in the same currency to the various platforms. By default, the tags should capture the same revenue data as displayed on the website. But sometimes users might set some currency conversions rules in the platforms which can lead to discrepancies.

This list is probably not comprehensive, but I have detailed out all the common discrepancy reasons that I have experienced over my 6 years in digital marketing. If you have tried all the above measures and more, and there is still no breakthrough, maybe it’s time to consider one last option

my vendor’s solution is broken and I need to look elsewhere.

The views expressed here are my own, and have no association with my employers, present or past.

 

You may also like...