Marketing You Can Measure
Are you able to tie your marketing investments directly to revenue? If your answer is no, my advice is: steer clear. Avoid what you can’t measure.
Many startups make the mistake of focusing their marketing on “brand awareness.” Brand awareness is a risky investment — it can be expensive, require a lot of time and is famously hard to measure. Luckily, we have data on our side. Advances in machine learning and analytics make it possible to pinpoint your target customers and market to them with precision. When done right, you know exactly who you are trying to attract and exactly how to attract them. Testing allows you to make decisions that are not based on “gut feeling” or intuition, and are hopefully unbiased. These decisions allow you to win, and that’s what we are all in this for, right, to win?
Understand what your customers want
Human nature says the best way to find out what customers want is simply to ask them. That leads startups to spend a ton of time and investment on focus groups and surveys, believing that the insights gained will tell them exactly what customers want. The problem is that customers usually do not know what they want and are unable to imagine things that are truly new — which is what most startups are trying to do. In other words, focus groups are really only good at building faster horses (and, no, Henry Ford did not say that).
Instead of asking customers what they want, measure what they actually do. Almost all marketing and sales interactions generate digital exhaust. Real exhaust smells gross; digital exhaust smells like future success. Formulate hypotheses, then test them in the market to measure customers’ actual behavior. Sometimes your hypothesis will turn out to be right. Great, scale it. Sometimes your hypothesis will be flat out wrong. Ditch it. Sometimes your hypothesis will be kind of right and kind of wrong. Fix it and iterate, then test again. Testing and data reveal what your customer really wants. Use the results to make your marketing better, and to continue to improve it.
Measure the channels that work
Data-driven marketing tells you which channels work and which channels don’t. Too often, our intuition about channels is not correct. Figure out how to measure the impact of your channels — but more than just conversion. For example, does your direct marketing lead to phone or web inquiries? You might make different support decisions depending on that answer. Once you have figured out how to measure impact, do a test run. Try multiple channels for reaching your customers and measure the effectiveness of each. Then, use the channels that work and leave behind the channels that don’t.
Reduce your customer acquisition costs
When you effectively test your marketing, you will be more able to acquire customers in the market, which will yield a lower acquisition cost. After you’ve tested and determined who to target, what they want, and how to reach them — measure cost of customer acquisition maniacally. Aim to cut customer acquisition costs on a monthly basis. Use data to continue to determine what’s working, ditching what’s not. When your cost of acquisition is lower than your competitors’, you can grow more rapidly. And win, remember?
This isn’t just talk. At ZestFinance we put as much technical effort into creating models for marketing as we put into our credit underwriting model. We use thousands of signals, and several machine learning models, to decide which channels to use for marketing and which customers (across the different channels) to reach out to. Our marketing approach is as critical to our lending business as our ability to underwrite loans. It enables us to reach the right customer at the right time with the right product, at lower cost.
Mr. Merrill is founder and CEO of ZestFinance and a former CIO and Vice President of Engineering at Google.
Originally published on wsj.com by Douglas Merrill