You want ROI? I'll show you ROI.

That’s how I started responding after the third client in a row asked me about payback. It’s a fair question, but let’s be honest: 90% of the time, it’s just a way for a bank to decline a new direction, so they don’t have to take on “grunt work”.

Because ROI is just the tip of the iceberg, beneath the surface lies discipline and hundreds of experiments.

Fifteen years ago, in the early days of CRM at PrivatBank, we thought we were the hotshots: we ran 10-15 campaigns a month, managed everything manually, and spent hours arguing about what would take off 100%. That’s how we thought everything was under control.

Till we met a consultant who had been developing CRM for a British bank for 15 years. He asked how many communications we were running. I easily gave him the number. And he replied: “I see, but until you’re running 100-200 campaigns a day and measuring conversion, your CRM isn’t working.”

It took me just one meeting to realize – we’re just arguing about CRM instead of running it.

Finding the optimal solution requires a system that makes testing hypotheses as simple as possible.

And I was hated for the first time in the bank.
Because I went to the team and said, “We are building a rapid testing system. From now on, you’ll input CRM messages by yourselves.”

Once the system launched and we learned how to automate message chains, I was hated for the second time. This time, I forced people to build marketing campaigns and examine segments. Simply make a test on 1,000 customers. If there’s conversion, we scale. And that’s it. The debates stopped because reality made the decisions.

Then, before asking about ROI, you have to look “under the water”, where the work is. A control group should be a bank standard to see how many people take action without any messages. For us, it was always around 1-2%.

After PrivatBank, I stayed away from CRM for a long time. But then it hit me that a cashback module is just a CRM. And it follows the same logic.

To make it work, you have to:
– Test different segments.
– Measure conversion against a control group.
– Reduce frictions to launch new experiments faster.
– Find a manager to run your system.

I know that people love results, but they are annoyed by the work required to get them 😉

I can draw up brilliant numbers for you in a financial model. But a model alone doesn’t guarantee you profit.
So, are you ready to be hated until you build the system that actually delivers ROI? 😉