One of the most important metrics to keep your eye on in business is your Customer Acquisition Cost (CAC). It tells you how much it's costing you to find each new customer that you get. But if your CAC is too high, just staring at that one number won't give you much insight into how to bring it down.
Finding cheaper ways to reach you customers involves some trial and error. You'll need to start by coming up with ideas for marketing experiments and then you'll need to measure the results of those experiments.
Generating Marketing Experiment Ideas
How to come up with marketing ideas is not my area of expertise. You could look at what your competitors are doing, what similar companies in other sectors are doing, or what channels are emerging due to new technology. Or you could dream up something totally original. Hiring a marketing expert will almost certainly get you better quality ideas, but whatever ideas you have, the next stage is the same: try it and see what happens. That's where I come in.
Evaluating a marketing experiment involves isolating how much you spent on that experiment, and measuring how much money it brought in. It's not always easy, but there are a few common reports that will help.
Marketing Channels
Measuring where your customers came from is often helpful. Ideally, you want a report that looks like this:
| Channel | Average CAC | Average LTV |
|---|---|---|
| Facebook Ads | £9.32 | £12.79 |
| Yellow Pages | £13.81 | £12.53 |
| Organic Online Search | £6.45 | £12.05 |
| Cold Calling | £65.37 | £132.83 |
When you look at a report like this, the actionable insight should be clear. Don't renew the Yellow Pages listing. But how do you build this?
You'll need to know how much was spent on each channel. This involves going through the marketing budget and assigning things to channels. Some things may not fit neatly into one channel. If so, your only option is to divide the expense evenly between all the channels it contributes to. The cost of building a website, for example, should probably be divided evenly between all channels. This gives us a total budget for each channel.
Next we must assign customers to channels. If you sell online this is a bit easier, as you can often detect where people came from. Failing that, you could ask a random sample of your customers how they heard about you or just try and make intelligent guesses. Once you've assigned customers to channels, divide the total channel budget by total number of customers to get each channel's CAC.
Calculating a customer lifetime value (LTV) per channel is important too. Without it, you'd probably conclude that the cold calling channel in the table above wasn't profitable. Once you have worked out a system for assigning a customer to a channel, work out the LTV for customers within that channel to get each channel's LTV.
Cohorts Analysis
Another way to look at the CAC and LTV would be to break things down based on when each customer was acquired. This would give you a table like this:
| Cohort | Average CAC | Average LTV |
|---|---|---|
| Jan 2013 | £14.51 | £20.93 |
| Feb 2013 | £14.81 | £21.53 |
| Mar 2013 | £19.45 | £33.05 |
| Apr 2013 | £15.37 | £24.83 |
What's clear from this table, is that whatever you did in Mar 2013 cost a bit more money, but brought in more profitable customers. So, whatever it was, do it again.
As long as you're recording something uniquely personal for each sale (such as an email address) then Cohort Analysis is easier than channel analysis because you know for sure when each customer was acquired and how much they've spent. The process is very similar though, assign each customer to a cohort, add up how much you spent on marketing in each time period and divide that by the number of customers acquired. Then work out the LTV for the customers in each cohort.
Marketing Campaign Breakdown
Probably the most detailed approach would be to break things down by campaign, so you get something like this:
| Campaign | Average CAC | Average LTV |
|---|---|---|
| Facebook - Women in London - Ad #1 | £21.31 | £30.93 |
| Facebook - Women in Manchester - Ad #1 | £20.95 | £29.32 |
| Facebook - Women in London - Ad #2 | £22.93 | £29.58 |
| Facebook - Women in Manchester - Ad #2 | £23.12 | £30.02 |
What's clear from this table is that ad #1 performs better than ad #2. So you should probably use ad #1 in favour of ad #2. This type of table is very difficult to work out though. It's probably only do-able via online advertising channels as they're the only ones that can track exactly which ad your customer saw. There's also a danger that you'll get overwhelmed by the number of different permutations. If you have 4 or 5 different channels, and a dozen different campaigns running on each, then you could end up with a monstrous report. If that happens, I recommend you switch to a report that summarises the best and worst performing campaigns, rather than a complete list.
Tying it Together
Underpinning all of this is a system that links where each customer came from, to what they bought when they arrived. This is not something that generally exists by default. I find even businesses that sell online tend to fail when it comes to linking their advertising systems (which will track which ads are clicked on) to link their sales systems (which tracks what people bought). The simplest way of doing this is to use the e-commerce tracking features in Google Analytics (you're using Google Analytics right?). But longer term, it's worth storing information about where your customers arrived from in your own databases, so they can be more easily linked to your sales.