7 Reasons Why Your Sales Don’t Match Google Analytics Transactions

7 Reasons Why Your Sales Don’t Match Google Analytics Transactions

Frustrated that your Google Analytics (GA) transactions do not match your actual sales?

Your digital analytics tools will almost always have a discrepancy with your actual sales. The very nature of digital analytics data is to be directional not accurate. This is because it is built on a framework of cookies. Cookies will never be 100% accurate for many reasons, including for example, people or plugins blocking or deleting them.

For GA to be directional, the data should be as close as reasonably possible. A reasonable expectation would be for the data to vary 1 – 10% between the two data systems or potentially higher if your website audience is more tech savvy. When you do notice a major discrepancy (>10% difference), try testing out the following to narrow in on the reasons why.

It is more common to see fewer transactions in GA than in the back-end. If this is your scenario consider it could be that:

1. GA is not capturing an add-on purchase path.

Purchase Path

In most cases, new purchase paths will not automatically be tracked. For example, let’s say a website that sells custom shoes decides to offer a training course on shoe making. This purchase path is built into the website so that customers can easily purchase the shoe making course online.

Since the payment path for courses is designed differently than a regular product purchase path, your eCommerce tracking code may not immediately capture this new method of purchase. When you look at your sales data you will find GA sales below actual sales because GA does not include course purchases.

To isolate for this discrepancy, consider cross-referencing transaction IDs of purchases not found in GA with the purchase path.

2. The information-per-hit limit is surpassed.

Movie theatre popcornWhen the amount of information being sent in a single hit surpasses 8KB, the hit will drop and not be passed to GA. Imagine you were buying movie tickets online for a group trip. If each person customized their order and added variations of popcorn, fountain drinks, 3D or regular viewings, skittles or M&M’s, etc. then each product transaction could contain a lot of information.

Try scanning your back-end sales data for high value purchases. Are these the only purchases missing from GA? If so, this could mean you are overloading the hit with too much information.

3. The hits-per-session limit is reached.

The per session hit limit is 500 hits. If you track a lot of interactions on your websiteOnline Shopping you may unknowingly be running into this limitation. For example, imagine you did all your Christmas shopping from one website. Say you bought different shirts, pants, picture frames, kitchenware, board games and books, in all different shapes, sizes, colours and brands. You searched by top-rated, most- to least- popular, lowest- to highest- price, read customer reviews, then decided what to buy. When an online customer buys in this way they can easily max out the per session hit limit.

To test for this issue, choose some of the purchase scenarios that did not work and simulate these purchase sessions to see if the transaction fires. If you are missing hits for this reason, you may want to re-evaluate what is important to be captured on your website. Just because you can track an interaction, does not mean you should.

4. The receipt page closes before your GA tag fires.

Phone ReceiptThis can happen for a number of reasons. Sometimes a customer may not give the page enough time to load before closing their browser. Other times the GA/GTM snippet is really low on the page and does not load before a customer navigates elsewhere. Since this hit did not send from your website, GA will not receive the transaction. This can lead to your transactions being understated in GA.

Check where your GA/GTM snippet is placed on your site. If it is in the header, it likely will have time to load. If not, try moving it to the header to see if this narrows the discrepancy.

Alternatively, you could be seeing more transactions in GA than actual sales. This could be caused by:

5. Multiple tags or containers.

multiple containers

If more than one of your containers are tracking  purchases or if your GTM tag/trigger configuration has gone awry, you could be sending transactions to GA multiple times, inflating total transactions.

Using the Google Tag Assistant plugin in Chrome, try a test transaction and record the session. If multiple containers or tags are firing for the same transaction this will be captured in the recording. Another high-level check would be if the value of certain transactions is double or triple what you would expect from back-end sales.

6. Receipt pages not expiring.

page-timeout

For example, if someone makes a purchase and leaves the receipt page open on their browser (or bookmarks it) then comes back to the tab later, they could send duplicate transactions to GA. Every time this person comes back to the page they will inflate your GA sales data.

Look in your GA data for purchases of the same value more than once by the same userID to confirm if this is the case.

7. GA including call centre purchases.

Call Center Rep

If your business has a call centre it could be the cause of your sales discrepancy. Sometimes these transactions are separated from online sales in your back-end reports. If these agents buy for your customers online through the same purchasing flows as customers then they would inflate your expected online sales.

For transactions you do not see in your backend reports, consider adding ‘city’ as a secondary dimension to see if the unexpected transactions are grouped in locations that match up with your call centre locations.

When in doubt, always go with your back-end eCommerce platform because it was designed with the intent of accurately measuring your day-to-day sales.

If you are experiencing any of the above or something different and need help narrowing in on the gap, message us and we can help you investigate the issue further.

By |2018-05-22T11:38:38+00:00May 16th, 2018|2 Comments

2 Comments

  1. Jane Ansara May 31, 2018 at 12:06 pm - Reply

    Completely agree. Am always setting the expectation that analytics data is for measuring website performance and should not be used the system of record for transactions or other such metrics.

    • Sarah Hendriks June 1, 2018 at 9:29 am - Reply

      Exactly! I think it is hard for management and analysts to accept less than 100% accuracy when business decisions are at stake. To me, I see digital analytics as both an art and a science. Part of being a good analyst is understanding the restrictions of the type of data we work with and being okay with it.

      Thanks for your comment Jane.

Leave A Comment