Brad Hargreaves | Building Things

Brad Hargreaves on entrepreneurship, community and life

This Is Not Your CAC

with 4 comments

In my brief time thus far as a VC, I’ve seen entrepreneurs repeatedly screw up one key metric again and again: customer acquisition cost (CAC).

Along with lifetime value (LTV), CAC is one of the most important metrics consumer-facing businesses need to watch. But so many of them get it wrong and do so in a very dangerous way.

The entrepreneurs who get it wrong calculate CAC as follows:

CAC = Advertising Spend / Total New Customers

Instead of:

CAC = Advertising Spend / Total New Customers Attributable to Paid Advertising

Counting all your customers in your CAC metric – rather than just those acquired through paid marketing – is playing with fire. As an entrepreneur, I learned that the worst possible result of a test is a false positive: a result that incorrectly makes a hypothesis look correct when it actually doesn’t work. False positives will waste time, money and precious focus.

Calculating your CAC as a “blend” of all customers is just asking to be on the receiving end of a false positive. Your paid campaigns could very well be unprofitable, but you are making them appear profitable by blending in customers who found your product through organic channels – referral, word-of-mouth, organic search, or social, to name a few.

This is a huge problem since traffic from those organic channels doesn’t scale linearly with traffic from paid channels as a paid advertising budget increases. If you are spending (say) $1,000 on paid marketing and getting 20 new customers per day, it does not follow that your CAC is $50. 5 of those customers may be from paid marketing campaigns while the other 15 are from organic sources, so your CAC is actually $200. When you scale your paid budget to $2,000, you won’t be getting 40 customers per day – only 25 at best. And that’s assuming that CAC stays constant as your budget increases, which is a very optimistic assumption.

Using “blended CAC” in a presentation is a very quick way to get me to pass on an investment. When I hear a pitch, I care less about whether or not a company’s initial paid marketing tests were profitable or not. There are numerous factors that can drive campaign profitability, and there are a lot of ways to acquire customers these days. Rather, I care about how an entrepreneur uses data and thinks about key metrics. Showing a blended CAC means an entrepreneur is loose with the numbers at best and deceptive at worst. And as a company grows and the numbers get bigger, either of those are very bad things.

Written by Brad Hargreaves

March 15th, 2015 at 6:48 am

  • John Jackovin

    Couldn’t agree more. Makes A/B testing REALLY difficult if you blend your results.

  • Ross Gordon

    Very interesting post, however one thing to consider is that early stage startups (and even many late-stage ones) lack the ability and resources to truly attribute CAC to specific channels. Even huge brands often struggle with attribution and some companies have entire resources dedicated to attribution modeling. For example, we may spend $2,000 in one day on Facebook but only see 10 direct click through conversions. But on the same day we may see a 15% bump in conversions come through branded paid search and another 20% increase come through organic search, and another 5 sales come in through the phone, 10 come in from other websites (where people landed after searching our brand) and 5 more come in through retargeting a week later. As you can see, the CAC Facebook spend is not really $200, although basic analytics may make it seem that way. It gets even more complex for companies that advertise on TV, YouTube, Print, Radio, etc. So I think blended CAC is still an important metric to look at the overall performance of CAC/LTV, especially for early state startups that lack the resources to do sophisticated attribution modeling.

  • MikePugh

    You really need to do both: track (1) campaign level metrics and (2) aggregate level metrics. Tracking only at the campaign level ignores branding benefits, lift across channels, and the difficulty of measuring some campaigns. Measuring only at the gross level covers up poor performers and hides good ones. Doing both still leaves plenty of questions unanswered, but at least you’re looking from different angles when making tough judgment calls.

  • Alex Chaidaroglou

    A very interesting post and I totally agree with this. I think if early stage startups stick to more usual methods of PPC it won’t be very hard to track and especially with all the new tools out there.

    I believe it’s even better to track CAC and LTV per advertising channel, to be fully transparent.