Cost per Acquisition (CPA) – Not all are Created Equal!
A few years ago, I was working with a client to optimize their digital new-customer acquisition campaign. They were bringing in enormous volumes of customers, spending well into double digit millions, yet frustrated with their return on investment. Similarly, another client was spending millions per month in advertising and they had an enormous response. Was the latter a success and the former a failure?
If you’re using Cost per Acquisition (also known as Cost per Add/CPA) as your only metric to analyze the campaign, then yes, the second example was a smashing success. It drove volumes to the site and enrollments grew. The CPA was extremely low and everyone was happy. But they shouldn’t necessarily have been celebrating yet. What about the quality? Measuring the campaign on a CPA basis seemed right, but it was not assessing value. CPA alone isn’t enough. You need to measure quality effectiveness by cost-per-effective-add.
Effectiveness can be measured by the quality of the new customer engagement – are they responsive? Is their feedback truthful, accurate, complete, or unbiased? If engagement is high, you may be willing to pay more and have a higher CPA. Or are they gaming you and motivated by the incentive offer? What about quantity? At a $1.00 CPA, you might need a customer to engage just once. But at $10.00, I’d need it 10x just to have the same breakeven. This is overly simplified, but too often, this is all many companies are doing in terms of evaluating whether a digital acquisition campaign worked.
What’s really missing, is taking a longer-term view of the event. Here we are, several years later and I am surprised that so many businesses still use standard Click-thru Rate (CTR) and CPA to determine whether a source is viable. While this is effective for a top line, you really need to marry this with lifetime value (LTV) of customers by source. This allows you to understand the value of using a more costly source or channel over a longer period.
At the other end of the spectrum, many clients are seeking to leverage media mix modeling to optimize their allocations and marketing budgets. But surprisingly, many use premature inputs – meaning a focus on response and CPA and not building a learning loop of including LTV by source inputs. This missing piece, derived over time, helps clients to truly determine their source and channel optimization mix.
I have a personal bias of looking at acquisition sources (digital, email list sources) by LTV combined with look-alike modeling to determine which sources bring in my best LTV customers, and at what acquisition price point. Stable sources allow for predictability in the planning forecast.
Direct mail or even email may be more costly channels, but if they are bringing in new loyal customers who spend significantly more over time, their ROI would justify the higher marketing spend.
And while direct mail is a surprising channel to advocate today, think about this – is your personal inbox flooded and ignored and have you noticed a marked decline in direct mail at your home? Between my shifting most behaviors online, I barely get direct mail. And when I do, I look at and use the offers that I would have deleted via email overload.
And that is an effective outcome!
Pamela Veraart is SVP, Head of Strategy at C2G Partners.