Ecommerce brands are addicted to sitewide offers. But they don’t have to be. AI for ecommerce can empower online retailers to pull back on their promos without sacrificing revenue by targeting discounts to a particular segment of visitors who need an incentive to purchase. Using AI to score millions of site visitors by their likelihood to purchase uncovered that each segment reacts differently to offers.
- Likely Buyers: this is a group of motivated buyers who would buy from you regardless of being offered a discount. While sitewide offers have a minimal impact on this group, we have found that many brands spend as much as 75% of their coupon budget on such transactions.
- Unlikely Buyers: this group is not on your site to shop, and they will most likely not buy regardless of whether or not a discount is presented. Exposing sitewide offers to this group can dilute your brand image and even train these site visitors to come looking for discounts next time.
- Influenceable Buyers (On-the-Fence Shoppers): this is a group of visitors who are considering a purchase, but are not yet convinced. They might need to compare prices or shipping options with competitive sites. Sitewide offers are effective for this group, but the offers could be presented much more effectively by using targeted offers with single-use promo codes (SUPCs) and a short-term expiration so that the shopper knows they are getting a good deal with some urgency to transact before it’s too late.
Session AI in-session marketing identifies all three segments of shoppers early enough in their visit to tailor offer deployment or suppression and maximize your conversion rate and margin. This solution is called Session AI real-time offers.
How real-time offers work
Session AI’s patented Customer DNA™ consumes clickstream data from your ecommerce site to predict whether each session will result in a purchase — within five clicks. We do not require 3rd party or historical data, so the solution works well for anonymous, first-time site visitors. When our prediction indicates that a shopper is on-the-fence, we trigger actions to influence that session before the visitor leaves the site. A real-time offer is a targeted discount to the on-the-fence shopper that lifts the conversion rate by giving a well-timed nudge to complete their purchase.
Learn more about the value of real-time offers, here.
How do we know that real-time offers are better than sitewide offers? Isn’t it true that if I give an offer, you will be more likely to buy? We have tested this question on billions of visits on top-tier ecommerce sites, and the answer is “Not necessarily.”
In fact, it is counterproductive to give offers to some visitors!
Here is an example test result: We used our Session AI EPP (early purchase prediction) model to assign a purchase propensity score for each website visitor. We then gave the same offer to all visitors in a test group, regardless of their purchase propensity. We measured lift over the control for each propensity group to determine the effectiveness of real-time offers. We have done this for several of our customers and the results are consistent.
- Conversion rate lift for on-the-fence shoppers is significantly higher than for all other groups.
- The increase in conversion rate for some groups was barely worth the impact on margin.
- The conversion rate for some groups is actually reduced(!) by showing offers.
Great, real-time offers work. But my promo calendar is completely saturated with other offers. How can I roll that back to introduce real-time offers instead?
Three techniques to transition from sitewide offers to real-time offers
If you already feel the pain from thin margins and excessive promo spending, then you might want to skip directly to Technique #3. But if you are like most ecommerce leaders, you may need to experiment with the first two techniques to prove this works on your site before involving the broader team.
Technique #1: Real-time offers as sweeteners.
This can be the fastest path to a quick win as it does not require coordination with your other promos. The way it works is that you simply come up with a new offer that can stack on top of your existing promos. Then issue that offer to the on-the-fence shoppers and measure the lift over a control group.
Here are some example offer types that might work for your organization.
- If your offer system does not support multiple promo codes (stacked promos), then you could offer an alternative, more appealing promo for the on-the-fence shoppers. For example, if you are running a seasonal 10% sitewide offer, you could show a message like this: “Take
10%15% off in the next 30 minutes with offer code SummerDeals 123456″ (where 123456 is the single-use promo code or SUPC).
- If you can stack offers but are concerned that an additional discount will be too noisy, you could use a non-monetary offer such as Free Expedited Shipping on top of existing discounts. This could be particularly effective if you don’t normally do free shipping.
- If you have a GWP (Gift With Purchase) program and available gift inventory, then you could try running it only for the on-the-fence shoppers.
Pros of this technique: Does not require significant coordination with other teams. Typically provides net incremental profit compared to the control group. Does not require fixed scheduling.
Cons of this technique: Requires incremental offer budget. Temporarily increases offer exposure to a small group.
Technique #2: Use a pre-promo (or post-promo) to test real-time offers
Suppose you have an event on your promo calendar that starts on Feb 15. Your finance team has already analyzed margins and given you the go-ahead. You have a coupon code and your ad creative is ready. You can test real-time offers by silently starting early for on-the-fence shoppers. This is a nice, clean way to test real-time offers without disrupting your promo calendar. Afterward, you can conduct a holistic analysis to compare the two techniques.
Note: If your upcoming promo does not use single-use promo codes, then we would recommend creating a batch of SUPCs that has the same discount value as the upcoming promotion in order to avoid exposing a code (which could be shared online).
Also, if you are concerned about the risk of accidentally exposing the promo code too early by mistake, you could run this as a post-promo — which is to silently run the same campaign after the scheduled time instead of before.
Pros of this technique: Does not disrupt your promo calendar. Piggybacks on existing margin analysis.
Cons of this technique: Inflexible timing constraints (must be attached to a scheduled promo). May require additional promo budget or you might need to borrow some from the scheduled promo.
Technique #3: Replace an existing promo
This technique is most similar to the long-term strategy that you will want to execute. In this technique, you remove an upcoming promo from your calendar. Don’t advertise the promo and avoid showing it sitewide. Instead, use the entire offer budget on visitors when they are identified as on-the-fence shoppers. This will drive a higher conversion rate lift and increase margins. Since you are not advertising this promo, it frees up advertising dollars for branding or retargeting. Then conduct a holistic profitability analysis. As you discover that this approach is financially effective, you will gain better control over your promotions, improve your brand image, and provide a better customer experience.
Pros of this technique: Improves margin by suppressing offers to those who don’t need it. Boosts brand value image by reducing offer exposure. Increases conversion rate for on-the-fence shoppers. Frees up ad dollars for other, unrelated campaigns. May enhance the quality of your inbound site traffic (fewer deal seekers).
Cons of this technique: Requires changes to your promo calendar. May impact inbound site traffic volume. You might need to convince stakeholders in your organization.
Putting it all together, here is an example of a 3-phase roadmap to introduce real-time offers.