Why CPA is an outdated metric to measure success

Related Articles


We are RocketMill: This is Izzy Scott-Evans


We are RocketMill: This is Katy Kaminarides


Why CPA is an outdated metric to measure success

Working in Paid Media, I have conversations about CPA, CPL and direct ROI on a daily basis. But the world is changing, and companies which focus solely on last-click CPA as a measure of success will soon find themselves out-competed by smarter brands taking a holistic view of their sales funnel.

Think of it this way: if you’re anything like me, you’re much more likely to buy something from a brand that has a unique voice, and a sense of purpose behind it. I’m far from alone. The web is absolutely drowning in no-name grey market goods from companies you’ve never heard of, at prices that seem ludicrously low. Most established western companies simply can’t compete on price, so they need to find another battleground. Can you put a price on a set of values?

Patagonia is perhaps my favourite example of a value-driven company. It believes so wholeheartedly in the value of public lands that it spent close to $1 million on a TV campaign asking people not to buy clothes, but to donate to a campaign to stop the Trump administration selling off National Parks. Patagonia state its mission as a company is to “save our home planet”, and that comes across in every aspect of its communication. It’s not just the National Parks campaign I mention in my talk, it’s giving the windfall from a tax cut to environmental charities, or shining a spotlight on the issues caused by hydroelectric plants.

Patagonia is a cause-led business – a cause that most people will agree with – which gives people a genuine reason to choose its products over whatever pops up in a lowest-price-first Amazon search for puffer jackets.

It’s hard to imagine a situation where Patagonia’s CMO sends an email asking for a return on investment on every acre of forest they’ve saved; yet all that activity undeniably has an impact further down the funnel.

Thinking beyond last-click attribution

Data-driven attribution models have become increasingly popular in recent years, and it’s obvious to see why – they offer the clearest picture yet of how every piece of your marketing pie comes together to generate sales. Unfortunately developing and deploying a data-driven attribution model still requires a fair amount of technical knowledge (if you’re looking to develop your own, our Data & Insight team are a great place to start), but Google has introduced some fantastic tools to bog standard Google Analytics accounts that let you explore how different models compare when attributing sales.

The Assisted Conversions Report

Found under Conversions → Multi-Channel Funnels, this report will show you which channels play a larger role further up the funnel. It’s not hugely precise (an assisted conversion simply means a channel played a part; the report doesn’t give you any data on how it factored into an individual conversion. Also revenue is stacked, so make sure you’re not double-counting!), but gives you some great insight into channels that play a highly assistive role, but may not get much credit using a purely last-click model.

The Top Conversion Paths Report

This report allows you to go one level deeper. You can also find it under Conversions → Multi-Channel Funnels, and it allows you to explore what combinations of channels have lead to the most sales. If you make good use of UTM parameters, you can even attribute down to a campaign level. For example, if you ran a big email campaign you might find a small number of conversions straight away. On the face of it this number might be a bit underwhelming, but this report will show you users that come back through Direct or Paid channels later on and convert, allowing you to properly measure the impact of that activity.

The Model Comparison Tool

Finally, this tool is a great way to compare last-click against other pre-built models within Google Analytics. Found under Conversions → Attribution, you’re able to easy compare channel groupings and source/medium combinations across different attribution models such as first-interaction, linear and data-driven. This is one of my favourite reports, as you’re able to easily identify the percentage difference in conversion reporting across the various models, giving you the information needed to make better decisions on your branding activity.

CPAs are still useful in the right place

Don’t get me wrong, I’m not going to stop reporting on CPA any time soon – but only where it’s a relevant and useful metric. If you’re running activity that isn’t directly acquisition-focused, consider putting in softer goals that better represent the aim of the campaign. Once you’re measuring your activity against the right metrics, use the reports above – or your own data-driven attribution model – to work out whether it’s a useful addition to your marketing mix.