4 issues you face with Adobe Attribution IQ and how to fix them

You understood the importance of marketing attribution, and wanted to do more than last touch attribution. Yet, you faced challenges while building attribution model using Adobe Attribution IQ tool.

Sound familiar? You’re not alone.

Working with Adobe workspace can sometimes be challenging, and the lack of tutorials online does not make it any easier.

It is not so much of the UI, but more of the fundamental understanding of data and the way Adobe build its components.

We will be covering 4 potential issues you might face with Adobe Attribution IQ, and provide you with the knowledge to make attribution work.

1. Showing different numbers

One of the easiest way to check if you are doing it right, is to see the total number for the metric using attribution model.

As you can see above, the numbers are not the same. That should not be the case. The figure in the purple box should have been 15,778 instead of 16,267.

Attribution model does not change the total number.

This is a classic example of Adobe Workspace component misuse, and it is more common than you would think.

Let’s focus on the shoe component (yellow) on the right as shown above. It is created by dragging and dropping the value “shoe” from the dimension variable that captures “Product Type”.

If you have been doing this instead of building a proper segment, STOP NOW.

The shoe component does not represent a segment at all, and should only be used for building a segment.

Segment can represent a hit, a visit or a unique visitor. But the shoe component represents nothing at all on its own.

Therefore, if you are looking to pair segment with Attribution models, all you need to do is to create proper segments and apply it as such.

2. Metrics cannot be used for Attribution modelling

If you are having the issue above where the Attribution modelling option is greyed out, you are likely using the wrong metrics.

Adobe Workspace allows us to slice and dice data, but that doesn’t mean we should be doing that aimlessly. In many cases, the results are misrepresented and do not make any sense.

For attribution modelling using Adobe Attribution IQ, we recommend you to only use custom metrics you have created.

Below is a list of metrics that are not supported for Adobe Attribution IQ:

  • Unique Visitors
  • Visits
  • Occurences
  • Page Views
  • A4T metrics
  • Time Spent metrics
  • Bounces
  • Bounce Rate
  • Entries
  • Exits
  • Pages Not Found
  • Searches
  • Single Page Visits
  • Single Access

You can find the same list at Adobe Attribution FAQ page.

The list is made up of default metrics provided by Adobe, and they cannot be used to enable attribution modelling.

“But wait.. page views too??”

Yes. You are expected to capture all meaningful user actions and conversions using custom events.

If it is not done yet, talk to your web analytics consultant on fixing it asap. This is the prerequisite for consistent and accurate reporting. Attribution modelling just happened to surface the bad practice of using page views for conversion tracking.

This is what it looks like when you ticked the checkbox “Use non-default attribution model” for an eligible metric.

3. Numbers not adding up

Apart from the number issue in point 1, there might be another “number discrepency” plaguing you.

This is pretty straightforward if you are familiar with the different kinds of attribution models.

As shown in the purple box, if you add up all the values in that column, it will definitely be larger than the total value shown (15,778).

Reason being, that attribution model is Participation.

Using the Participation attribution model, it gives 100% to all unique touch points, hence the inflated total number. But it does deduplication channels in its calculation.

Example

  • Bob saw a display ad on Facebook and clicked on it. He did not take any further action.
  • 4 days later, Bob saw the same Facebook ad again and clicked on it. Still, he did not take any further action.
  • 2 weeks later, remembering the Facebook ad, Bob searched for the product online and clicked on a SEM ad. This time, he made a purchase.

Using Last touch attribution, 100% of the credit goes to SEM ad.

  • Facebook ad: 0
  • SEM ad: 1

Using First touch attribution, 100% of the credit goes to Facebook ad.

  • Facebook ad: 1
  • SEM ad: 0

Using Linear attribution, 1/3 of the credit goes to each of the 3 touch points.

  • Facebook ad: 2/3
  • SEM ad: 1/3

Using Participation attribution, 100% of the credit goes to each of the 2 unique touch points.

  • Facebook ad: 1
  • SEM ad: 1

Every attribution model tells a different story. Plan wisely to avoid data bias.

4. Calculated metrics cannot be used

In some web analytics setup, consultants make use of calculated metrics to generate segment-specific metrics.

If you are trying to enable attribution modelling for such calculated metrics, you will face the same issue mentioned in point 2. The option will be greyed out.

Enabling attribution modelling for calculated metrics is still possible.

Steps to enable attribution modelling for calculated metrics:

Edit the calculated metric. Click on the cog and tick the checkbox “Use non-default attribution model”.

Choose the attribution model from the dropbox. Click on “Apply” and you’re done.

With that, you can use the attribution-enabled calculated metric in your dashboard.

Conclusion

I have covered 4 of the more common issues you will face with Adobe Attribution IQ. If you are facing any other issue, feel free to drop me a comment and I will try my best to help you out.

Remember, a tool is only as strong as its user.

If you find this article useful, you can find more on https://webanalyticsguy.com .

About admin 32 Articles
Analytics Consultant who loves numbers and data! In love with Adobe Analytics.

Be the first to comment

Leave a Reply

Your email address will not be published.


*