Search the website

Can a Single Cross-Sell Campaign Change Customer Behavior?

Two recent campaigns created a sustained shift in customer behavior, proving the power of cross-selling

Cross-selling a customer to purchase from a more profitable category is one of marketing’s holy grails. Some products create a higher predicted customer lifetime value: a customer who buys only sourdough bread at your deli, but then switches over to wine, will grow your business. Rather than a one-time event, this is a change in customer behavior that has implications larger than a single campaign response.

In gaming, a very common cross-sell is from sports betting to casino, since in most cases the predicted LTV of a casino player or a multi-player (i.e., one playing both sports and casino) is up to five times higher than that of a sports-only player. A campaign for cross-selling sports players into casino games will usually offer a bonus or free currency for casino games, in an effort to introduce this new experience, with the long-term aim of migrating the player to this more valuable category. Similarly, e-commerce brands will offer discounts or create FOMO campaigns in order to create similar shifts.

But since this strategy is aimed at changing long-term customer behavior, its effects can’t be measured by the response from a single campaign alone. After designing the campaign’s target segment, offer, copy, channels and timing, the crucial final parameter to define is the campaign’s success metric.

Cross-sell campaigns should be evaluated based on two KPIs with different time frames:

• In the short term, the KPI is a statistically significant campaign response (e.g., bets placed or purchases made) by the campaign’s recipients, as compared with that of the campaign’s control group.

• In the long term, we will need to define a time period within which we can expect to see recurring behavior – this varies widely between industries, from a period of few days to a few months – and to again measure the campaign’s recipient group activity against that of the control group.

Measuring the campaign using only the short-term KPI might create a false impression that the campaign hasn’t been successful, since real-money bonuses or discount campaigns usually show a lower ROI for the campaign’s recipient group as compared with the control group. The result might even persuade the marketer to kill the campaign altogether. However, over the long term, the campaign might demonstrate a significantly higher ROI among the campaign’s recipient group versus its control group.

Sustained Change

We analyzed a cross-sell campaign, similar to the gaming example above, that targeted more than 10K players. The offer was a Free Bet at any of the casino games, and the intention of the campaign was to try to change customers’ behavior and to push them into playing in other, more valuable categories.

The short-term response rate for the players who received the campaign was 360% higher than that of the control group (which consisted of players with similar behaviors). This outstanding uplift is no surprise: the campaign offered players free money.


But the real question was whether some of the players would shift over from sports to casino (or to sports and casino) as their main playing category for the long term. For this campaign, the long-term period was set at two weeks – the period in which a casino player is expected to interact with the operator multiple times.

The result was a 90% increase in product change among the campaign’s recipients as compared with the control group: 90% more players who received the campaign started playing at the casino, and continued to do so over an extended period of time, as compared with sports players who didn’t receive the campaign.


A New Favorite Category

In e-commerce, cross-selling usually involves shifting the customer to a more profitable department or category, as in the case of pushing a customer who typically purchases leather belts to buy a leather coat. Cross-selling is tougher in this vertical and usually requires further research using techniques such as market basket analysis.

We analyzed an e-commerce cross-sell campaign that targeted customers who primarily shopped in the chocolates & sweets department, in terms of amount spent and number of items. The intent of this campaign was to try to change customers’ behavior and to push them into buying products from more valuable departments.

11K customers received the campaign, alongside a control group of similar customers who did not. After the long-term period of one month, 31% more of the campaign’s recipients had changed their favorite department to a higher-priced product category, such as wines & spirits, personal care & beauty or vitamins & minerals (depending on which particular offers they each received).


Although the response is less striking than that of the gaming campaign example (it takes a lot to change customer behavior patterns in e-commerce), a statistically significant higher percentage of customers who received the campaign changed their favorite department.
Note that it’s not only that the campaign successfully pushed customers into purchasing additional products, but that it actually changed the customers’ long-term behavior, causing them to purchase more from higher-value departments than their previous favorites.

A Valuable Shift

Both campaigns provide examples of the shift in customer behavior that may be produced by a single cross-sell campaign. Though a bit more difficult to measure and validate than regular campaigns, when a cross-sell campaign is analyzed correctly, it frequently shows a positive long-term change in customer LTV.

Published on

Omer Liss

Omer Liss is Optimove's Director of Strategic Services, helping CRM executives of top online entertainment and e-commerce businesses optimize their customer retention strategy. Omer has vast experience consulting clients, analyzing their customer data and revealing actionable, data-driven marketing insights. Omer holds a BSc in Industrial Engineering and Management, specializing in Information Systems.