How much of a discount does it take to get someone to order? Most operators guess, and usually guess high. The PizzaSMARTS probe mailing answered the question before the money was spent.
From the 1997 manual
“How would you like to know the results of your mailing before you make the big investment?… A probe mailing is a small test mailing consisting of different mailing offers.… Since these various probe offers hit your delivery area all at the same time you will know that the only factor affecting results is the message or coupon offer itself.”
Probe Mailings, Measuring Results
Two reasons to probe
The manual named two advantages, and both still hold:
- It cuts coupon cost. “The probe will tell you exactly how much coupon incentive is necessary to trigger a purchase.” If $2 off brings in nearly as many orders as $5 off, every order in the big campaign keeps $3 more.
- It makes the big campaign succeed. The probe shows what it takes to win over a new customer, a price point or an idea, before you commit the full budget.
How to do it today
- Pick two to five offers. Different discount levels, or different ideas: a free item, a combo, a dollar amount off.
- Send them at the same time to a small random sample. Mixing them in one send is what makes the test fair: same day, same weather, same competition.
- Give every offer its own code. A unique coupon code or link for each version tells you exactly which offer each order came from.
- Count redemptions and profit, not just response. The best offer is the one that makes the most money per piece sent, which is often not the one with the most redemptions.
- Roll out the winner. Then use a holdout group on the full campaign to confirm it.
Email, text and online ads make probing nearly free: split the list or the ad budget into versions and let them run for a few days. Postcards cost more to test, which is exactly why a small probe before a large mailing pays for itself.
What to measure
For each offer: response rate, and profit per piece sent after the discount and food cost. Choose the offer with the best profit, not the most coupons returned.
Every one of these tests needs a number to judge by. The next lesson is about choosing the right one: one number per campaign.