The most common way to judge a marketing campaign is to look at total sales. It’s also the most misleading. Steve Green’s 1997 article “Steering Wheel or Divining Rod?” tells the story of a campaign that worked and looked like a failure.
From the 1997 manual
“I once had a client who decided that the mailing he did was a flop because his sales didn’t go up.… The mailing he did was an especially attractive offer designed to recruit new customers. It was sent only to non-customers and resulted in three times the normal rate of new customers for the month. The mailing was obviously successful with the group that it reached, however since regular customer sales dropped at the same time of the mailing, it looked as though the mailing didn’t work.”
How do you Market? Steering Wheel or Divining Rod?, Measuring Results
Without the breakdown, that client would have scrapped an excellent campaign. The manual’s point: “With measurable results favorable and unfavorable information can be used like a steering wheel.” Without them, you’re left with gut feel, the old marketing divining rod. And the manual reckoned that savvy marketers who check their data find their assumptions are wrong about 20% of the time.
Four campaigns, four numbers
Each of the four PizzaSMARTS targets has its own number, and a campaign aimed at that target should be judged by that number alone:
| If you market to… | …watch this number | How to calculate it |
|---|---|---|
| Regular customers | Purchase frequency | Orders this month ÷ customers who ordered this month |
| Non-customers | Market penetration | Active customers ÷ households in your area |
| First-time customers | Retention rate | Last month’s first-timers who ordered again ÷ last month’s first-timers |
| Fading customers | Rescue rate | Last month’s C customers who ordered ÷ last month’s C customers |
The manual named the two most important: market penetration and purchase frequency, the two halves of every pizzeria’s sales. It said they “should be to marketing, what Food and Labor statistics are to operations.”
How to do it today
- Calculate the four numbers every month. They come straight from your order data once your customers are graded.
- Write down the target before the campaign. Which number is this campaign supposed to move, and by how much?
- Judge it by that number. A non-customer campaign that triples new customers is a success, even in a month when regulars ordered less.
- Use a control group when you can. A holdout separates the campaign’s effect from everything else happening that month.
- Keep a monthly history. The manual’s “Rear View Mirror” report lined up six months side by side, so trends showed up before they reached total sales.
The one-page monthly review
Purchase frequency, market penetration, first-timer retention, rescue rate, plus the count of customers in each grade and the customer GPA. Six numbers, once a month, compared with last month. That’s the dashboard PizzaSMARTS mailed to its clients, and it still fits on one page.
Market penetration has one more use: it shows you where to look for new customers. That’s the battle map.