Scale up your online sales team quickly to cover your schedule during Covid-19. Contact us to get started

Serving the underserved: the path to incremental revenue measurement

4 key steps to successfully attaining incremental revenue through performance attribution

In our series “Serving the Underserved,” we have been discussing the untapped opportunity for e-commerce businesses to convert more visitors and drive up to 20% more revenue by providing real-time, live sales assistance to the right visitors, when the digital experience alone isn’t enough.

To recap, we identified 4 main considerations that are key to your success in “Serving the Underserved” and realizing the significant untapped revenue opportunity from these visitors.

1) Visitor Identification: Given the high volume of visitors and activity on your e-commerce site, does the solution use historical and real-time data to deliver rich predictive analytics to accurately identify those who you can profitably engage?

2) Digital Engagement:  Does the solution deliver a deep enough level of micro-segmentation so that you can accurately and effectively personalize your message and engagement approach to those prospective buyers in a way that enhances the customer experience?

3) Live Engagement: What kind of visitor engagement capabilities are available to you? You need strong sales agents who can engage visitors in whatever mode the visitor prefers and the ability to seamlessly shift engagement modes in the same session.

4) And finally, attaining incremental revenue through Performance Attribution is the last and most important consideration in being able to convert more visitors on your e-commerce site and the topic of this blog.

Increasing incremental revenue is critical for overall revenue growth. Yet practices for precisely and reliably measuring true incremental revenue are still maturing. Many of these have built-in inaccuracies or biases, such as neglecting to measure against a ‘like’ control group, misallocating users by segment, and basing decisions on unclean or unreconciled data. Ultimately, when companies are paying for incremental revenue lift, they must be confident the uplift is measured precisely and accurately.

Therefore, as you explore solutions, you will want to ensure that they can deliver real-time tracking of engagement performance & measurement of results at a micro-segmentation level, as well as the ability to leverage that information to continuously optimize all aspects of the program.

In our experience, there are 4 key steps for ensuring the accurate measurement of incremental revenue through performance attribution:


Dirty or “bad” data is out-of-date, redundant, duplicate, or incomplete information that resides on multiple IT systems across a company’s data center. According to Ovum Research, bad data can cost a business at least 30% of revenues. Gartner recently blamed poor data quality as a major reason 40% of all business initiatives fail to achieve targets.

In the highly competitive market of e-commerce, it goes without saying that the rising cost of bad data can’t be ignored. To build a proper foundation for accurate measurement, robust data cleansing must be an integral part of your business process. The ideal solution should ensure all visitors’ online behaviors and activities are tightly tagged and tracked – from starting point to final action.


We covered this extensively in our post on Website Visitor Qualification but reading a visitor’s ‘signals of intent’ – accurately and instantly – is now at the heart of conversion optimization for businesses engaged in digital commerce. These signals of intent are communicated through digital activity. The more attributes and activities you can identify, the more relevant and personalized the experience can be for visitors. And the more relevant and personalized the experience, the better the result – both for visitors and your company’s topline revenue. Scoring and profiling is your foundation for segmentation to provide precise personalization.


Many digital marketing tests use a randomly selected control group without first establishing segments of like visitors. The problem with this approach is that there is no way to guarantee the composition of test and control groups will equally represent the different types of visitors. The result? Potentially highly inaccurate read on results.  Such a significant miscalculation can have costly implications if you then roll out what you think is a winning test when in fact it may be the loser.

Before being assigned to a control or test group, visitors with similar attributes must be grouped together.
The solution must:

  • Account for meaningful differences in your scoring and segmentation of your visitors — to ensure control and test groups are comprised of like visitors.
  • Randomly assign visitors within a particular segment to either the control or the test group.
  • Ensure your control and test groups meet appropriate standards for statistical significance.


The right way to measure incremental revenue? Compare results against a like control group, and then sum the results from each segment.

We have outlined two approaches (one correct, one incorrect) for measuring performance using the same baseline information to illustrate just how wrong you can get it if you don’t segment first.

Website traffic: 100,000 visitors per month.
Two major product categories

  • Product category A: Low AOV with high traffic and transaction volume
  • Product category B: High AOV with low traffic and transaction volume.
  • A/B test method: Random selection of 60% of traffic to “test” and 40% of traffic to “control”.
  • Combined revenue from total traffic during test window: $1.5M

Test vs Control Approach A: Product categories treated as separate segments.

Revenue uplift for Segments 1 and 2 are calculated separately, and then combined. This is the accurate way to measure incremental revenue. Segment 1 contains a significantly higher number of visitors purchasing at a lower RPV, resulting in significantly lower RPV than Segment 2. Incremental revenue for each segment has been calculated separately. The resulting incremental revenue from each segment has been combined to obtain an accurate account of total incremental revenue: $161,390. If you segment by meaningful differences, compare each segment to a like control group, and then total the incremental revenue, you are accurately measuring revenue uplift.

Test vs Control Approach B: All traffic combined

Segments 1 & 2 are combined before calculating revenue uplift. This is an inaccurate way to measure incremental revenue.  Here’s why: These two segments have been combined before calculating incremental revenue uplift, resulting in $447,850.  If one compares this to the accurate scenario in Approach A, there is a net difference of $286,460 from the accurate measurement from Scenario A.  If you are not segmenting accurately, then you are not accurately measuring revenue uplift.

Segments 1 and 2 represent the first of what would typically be a long list of segments. Imagine 25+ segments, each with their own control group. It can get complicated very quickly. So when considering a solution, be sure it can calculate incremental revenue for each segment first, and then add the resulting incremental revenue for each segment together.

In summary, these 4 steps are key to delivering a solid foundation of incremental revenue measurement, and supporting an organization geared for revenue growth. To be confidant in your ability to drive this incremental revenue, your customer engagement solution should deliver real-time tracking of engagement performance & measurement of results at a micro-segmentation level, as well as continuous optimization through performance attribution.

More Articles

Ready to work together?

Get started with your best-in-class live agents or turnkey chat solution today