Normalization of response rates and why it matters.
Many years ago I was involved with the launch of a new computer magazine. It was highly successful as a start-up and we eventually became the 10th largest computer magazine (when measured by paid subscribers) in the country.
However satisfying the growth factor was, one of the more interesting lessons I learned during this launch was in observing firsthand the normalization of response rates.
We launched this magazine with a reasonably plain subscription card. It was inserted into a standard #10 envelope and we mailed the subscription offer to hundreds of thousands of people. We received roughly a 20% order rate in return. Those of you in the direct marketing world know this is a very unusual response rate. Off the charts successful if I might say so.
When the responses came in we opened the envelopes, pulled out the order form and deposited the checks, cash and credit card payments. We also set aside every subscription card so we could count the response codes (source codes) printed on the bottom of each card.
As I sat at the table using a tick sheet, to manually tally each source code, I noticed each code came in batches or waves. Sometimes, I would have eight our twelve cards in a row with the same code and then nothing for the next fifty or one hundred for that particular code. However, over, the course of time, the responses normalized across all of the codes until we could see clearly which code performed best. And of course each code was matched back to which list segment it represented.
This normalization happened not only on the individual codes but also across the entire database. We had a few million records to work with and although the first mailing produced an order rate of 20% plus that rate of return did not hold. As we mailed increasingly larger quantities, over time, the order rate dropped until finally after mailing the entire file we were seeing order rates in the 2-5% range.
Why does this matter and why is it noteworthy today? The experience was brought to mind during a recent startup conversation. A founder has invented a new product and started making phone calls into businesses. He ran into roadblocks in the first segment and found very little success. He pivoted to a different segment and suddenly found a good deal of success – within roughly twenty hours of work. The product works great and the market is quite large – millions of business across the country as prospects… and the initial twenty hours of work suggest they can make money easily, even while acquiring customers.
Not so fast, says my forty year marketing brain. Very few companies make money during the acquisition phase… remember, responses and orders come in batches and waves… and the larger the sample size the more true the results…original source still matters whether by text message, email, phone, ecommerce or snail mail. And so it goes.
Marketing is a funny world and it is always fun to be right, but most of the time it is a scrappy challenging fight to win market share and gain loyal customers. Especially when entering a crowded space that has been in market and developing for many, many decades.
Customers come in different sizes and represent a range of values. Some are easy to win and represent very little profit. Others take more time, patience and effort. Getting it right straight of the gate is wonderful, but as you roll out campaigns don’t be surprised if things change. It’s important to know your acquisition costs, the value of customers and how you will serve them over the long haul. And by all means, be prepared to test, pivot and learn as you scale up a company.