Lesson 4 of 5 · WhatsApp Offers
The offer ladder
When you don't know yet which offer works, you don't have to guess. Start small with one group and go up one step a week, only for those who did not buy. At the end you know how much each customer needed to come back.
The offer ladder is a method created by the MeContaí team to find out how much of an offer each customer needs to come back. The same group gets R$ 10 off in week 1, and only those who did not buy get R$ 20 in week 2 and R$ 30 in week 3. A step can be a product instead. Each customer stops at the lowest step that was enough.
Updated on .
Before you start: remember the action?
Without looking at lesson 3, on the action and the coupon, say which action to use for delivery and the two limits of the coupon in Goomer. Then check.
The offer ladder method
For one group of customers, R$ 10 off. For those who did not buy, R$ 20 in the second week, and R$ 30 in the third. It can be a product instead of a discount. (method by the MeContaí team, summed up in the ladder cheat sheet)
The four rules of the method:
- The group stays the same from start to finish. Nobody joins halfway through.
- Whoever buys leaves the ladder. Only those who did not buy get the next step.
- One step a week. That gives people time to read, think and order.
- Change one thing at a time. If the step goes from R$ 10 to R$ 20, the headline, the action and the send day stay the same. Otherwise you can't tell what made the difference.
Why the ladder saves money
A customer who comes back with R$ 10 never gets to see the R$ 30. Each customer stops at the lowest step that was enough for them. If the business sent R$ 30 to everyone from the start, it would pay R$ 30 even to the customers who would have come back for R$ 10.
A product step instead of a discount
Instead of a discount, a step can be a product. At Hamburgueria da Praça, the fictional restaurant used throughout the training, that could be the house fries (Batata da Casa), a drink, a dessert, or a free Burger Kids. Two things to keep in mind:
- A product costs the business less than its menu price. Free fries are worth the menu price to the customer, but they only cost the business what it spends to make them. In the simulator, enter that cost, not the price.
- Don't mix the two ladders in the same group. If week 1 is R$ 10 and week 2 is the fries, you can't tell whether what mattered was the amount or the type of offer. Run a discount ladder in one group and a product ladder in another, and compare the two.
Try it: simulate a ladder
The numbers below are just an example. Change the values and see what happens; "Back to the example" undoes your changes. Your own ladder comes in the exercise, in a separate simulator. Look at the "Received" column: it shrinks every week, because those who bought have left.
How to read the result without fooling yourself
The ladder shows how many came back at each step. Before you say "R$ 20 sells more", check for three traps:
| Trap | What happens | What to do |
|---|---|---|
| The group gets harder | The customers who were easy to bring back already came back in week 1. Week 3 only reaches the hardest ones. | A lower rate in week 3 doesn't mean R$ 30 is worse. Compare the same week across two groups, not one week against another. |
| Time brings people back too | Some of the people who ordered in week 2 would have ordered anyway: it was payday, or a holiday. | Randomly pick about 10% of the group to get no message at all, with at least 30 to 50 customers (a rule of thumb used in this course). Whatever that slice buys is what would have come in on its own. |
| The customer learns to wait | In the MeContaí team's experience, if the ladder always goes up, regular customers notice that waiting pays more. | Use the ladder to find out. Once you find the step that sells, lock it into the Foodservice track. |
The slice that gets no message is what science calls a control group: a slice drawn beforehand that receives nothing, so you have something to compare against. (course glossary; Scientific control, Wikipedia)
Want to go deeper? Ladder or random split: which test to use
The weekly ladder mixes three things: the amount of the step, the week it went out, and the fact that only the hard customers are left. To find out which amount sells more, there is a cleaner way: split the group at random (this is what's called an A/B test).
| Random split, same week | Weekly ladder | |
|---|---|---|
| How | Split the group at random into 4 slices: R$ 10, R$ 20, R$ 30 and nothing. Everyone gets the message on the same day. | R$ 10 in week 1, R$ 20 in week 2 and R$ 30 in week 3, only for those who did not buy. |
| Good for | Finding the amount that sells the most at the lowest cost. | Winning back customers who didn't return with the smaller offer. |
| Watch out | Each slice needs enough people: at least 30 to 50 customers. | No more than 3 steps, and never announce that the offer will get better (MeContaí guidance). |
Comparing groups drawn at random at the same time is the basis of a trustworthy test (Kohavi, Tang & Xu, 2020). The figure of 30 to 50 customers is a rule of thumb used in this course, not taken from a source.
What the business has to honor (Brazilian consumer law)
With the ladder or the random split, every message you send is an offer the business has to honor while it is valid. A customer who got R$ 10 and shows up with it before the deadline gets the R$ 10. That is what articles 30 and 35 of Brazil's Consumer Defense Code (CDC) (in Portuguese) say.
One coupon per step in Goomer
In Goomer, create one coupon per step, each with its own code and link, valid for one week (the "disponível entre", or "available between", field). That way every Goomer order already tells you at which step the customer came back. (Goomer Help Center, Cupons, in Portuguese, read on November 18, 2026; action and coupon) And the track rule still applies: whoever ordered leaves.
Quiz
1. In week 2, who gets the R$ 20 offer?
Whoever buys leaves the ladder. That way nobody gets more than they needed.
2. Why does the ladder cost less than giving R$ 30 to everyone?
A customer who came back with R$ 10 cost R$ 10, not R$ 30. In the simulator, that shows up as the savings.
3. In week 3 the rate dropped. What does that say about the R$ 30?
Week 3 reaches people who didn't respond to anything before. To compare offers, compare two groups in the same week.
4. Week 1 is R$ 10 and week 2 is free fries. What's the problem?
Change one thing at a time. Run the product ladder in a separate group.
5. Why leave 10% of the group without any message?
That's the control group. Without it, you give the offer credit for an order that would have come anyway.
6. You want to know whether R$ 20 sells more than R$ 10. What's the cleanest test?
A random split on the same day: the only difference between the groups is the amount. Different weeks mix the amount with the week. The random split is under "Want to go deeper?", above.
Mixing in the earlier lessons
Review the offer-strength scale
- Strong: ordered only once, or the last order was more than 90 days ago.
- Light: ordered 3 or more times and the last order was 45 days ago or less.
- Medium: any other case (2 orders, or 3 or more with the last one between 46 and 90 days ago).
- Adjustment: if they haven't ordered for more than twice their usual interval, go up one level.
7. What offer should the ladder start with for the group that ordered only once?
The offer-strength scale (lesson 2) tells you where the ladder starts. A single order means strong, so the steps already start higher.
8. How do you tell in Goomer at which step the customer came back?
Each coupon is a step (lesson 3). The code on the order tells you which offer brought the customer back.
Exercise: your business's first ladder
Pick a group from the Foodservice track (suggestion: customers who ordered only once). Build three steps for one ladder, either discount or product. In the simulator, enter what each step costs the business and how many customers you think will come back. Write that guess down somewhere else, because the simulator only keeps the last numbers you typed. After the send, replace it with the real numbers and compare them with what you wrote down. Whatever you type is saved only in this browser.
To wrap up: send the MeContaí team on WhatsApp your three steps, the group you chose and whether you'll leave the 10% without a message. If it's a product ladder, also tell us what each product costs you.
Tomorrow
Without looking, name the four rules of the ladder and the three reading traps. Then check them in the ladder cheat sheet.
Training