By Greg Laugero, VP Strategy & Services, Matt Coers, Senior Consultant, Digital Architect, Rob Saunders, Sr. Analytics and Optimization Consultant, and Angela Tang, Business Analyst

For some reason, we’ve been fielding questions about tracking revenue with web analytics questions recently. Overwhelmingly, the questions go something like this, “Why don’t my web analytics revenue numbers match my accounting system?” This generally creates some anxiety as it often leads to questioning the overall accuracy of their web analytics data.

It’s important to keep this issue in the proper perspective. We strongly caution customers from trying to get a 100% match. Revenue tracked with web analytics tools will be directional and should not be treated as a source of record. If you need to report accurate revenue driven by a web channel, then you should be looking at a business intelligence (BI) solution that can combine different sources into an accurate visualization of the “truth” of your web-based revenue, order volume and other key performance metrics.

In this post, we’ll discuss the specific reasons why there will almost always be a gap between your accounting systems and your web analytics tools when it comes to measuring revenue. Our best practice recommendation is to shoot for 90% accuracy (a 10% gap) for your US-based revenue. For non-US revenue, 80% accuracy (20% gap) is going to be a good benchmark. Trying to make these gaps smaller is possible but typically not cost effective. Spend your money on a BI solution that can get you closer if you have solid business reasons for doing so.

Why they might be different
There are several reasons why this revenue measurement gap may exist.

Web analytics tools often do not capture cancellations and other changes to existing orders. Many companies do not allow these changes to be made online, which means the web channel won’t have any record of these changes. However, if you are allowing these changes on your site, you might not be tracking these actions with your web analytics solution.

Some users’ browsers have security plugins that block tracking beacons and cookies from being sent to the analytics system. Whenever this happens, these orders will not be counted. Additionally, JavaScript errors and other network glitches can block beacons from transmitting properly.

A poor implementation of tags can lead to inaccuracies due to data being aggregated improperly. Here are some examples:

-Your analytics solution might be capturing the wrong SKU.

-Sometimes certain configurations that modify the final price are not tracked and included in the revenue number for an order.

-We’ve even seen situations where double counting happens when a user can reload or revisit a confirmation page, and because the revenue is counted on that page the revenue is counted again. (You can solve this problem by including a transaction ID in the data you capture so that the analytics tool will track only unique transactions.)

Sometimes your web analytics solution is not set up to record all revenue elements in an order. For example, we have seen an instance recently where product prices were tracked correctly, but discounts weren’t included. We’ve even seen situations where the actual billed amount isn’t made available to tracking scripts.

Doing business internationally can impact accuracy. Here are a few examples:

-Some clients may have multiple tiers of tax that can impact product pricing. For example, value-added taxes (VAT) may be included in product pricing whereas sales taxes are not.  So, displayed prices versus what the client considers the product price can be different based on the customer’s location.

-Currency conversion – web analytics tools may calculate the conversions different than your source of record.

-Some of the beacon tacking issues (due to browsers and cookies) we discussed above can be bigger issues outside the US.

Numeric’s Recommendations for Reconciling the Differences

Now that we understand the problems, it’s possible to lay out a method for closing the gap. That is, if you want to do it. Again, shoot for 90% for US-based revenue. If you’re below that, then you should do some troubleshooting using the challenges we cite above. Above 90% you’ll have a diminishing rate of return to try to close the gap.
To minimize discrepancies and get to that 90% or above benchmark, follow these steps:

1.  The best place to start is understanding the gap in between how your source of record calculates revenue and how your analytics tool does the same thing. Document the gaps, including missing variables and aggregation steps, in your analytics solution.

2.  Determine the cost of capturing the new variables and fixing any aggregation issues and decide if it’s worth it.

3.  If you do implement these changes, run new reports and analyze the remaining gap. Determine your level of comfort with the remaining gap and whether or not it’s worth additional cost to close it.

4.  For reporting purposes, make some statistical adjustments to your web analytics-reported revenue data based on the historical difference between web metrics and the source of record. Be transparent about those adjustments when communicating to stakeholders.

Another option is to simply skip to step 4 if your revenue gap is historically consistent. Again, be transparent that you are making the adjustment.
The Bottom Line – If revenue accuracy is important consider a Business Intelligence solution

Web analytics solutions do some metrics reasonably well – such as conversion rates, traffic, page views, paths, campaign tracking, top converting content. Even these metrics are approximations – more so than ever would or should be tolerated in our accounting systems. There is simply too much imprecision in how this data is collected when compared to more tightly regulated and audited back-end accounting systems.
This imprecision is highlighted when we compare what our web analytics tools say about revenue with what our back-end systems say. Web analytics will always suffer when held to that standard.
That said, there may be very good reasons you want greater than 90% revenue accuracy in your reports. You may want to have a highly accurate picture of average order value, for instance. If this is the case, then we recommend working with the appropriate people in your organization to put together a BI and data visualization solution that uses the best sources to get the most accurate picture. In our experience this is almost always a better use of time, money and effort. It will lead to longer-term solutions that you can build on later.