Retargeting campaign ROAS or CPA targets should be significantly lower than prospecting campaign targets. Up to 50% lower. Unfortunately, I see brands failing to set differentiated ROAS targets all the time.

As usual, I’m going to return to my favorite hobby horse: emphasizing that you need to be thinking about everything in your business on a marginal, or incremental, basis.

Incrementality explains why prospecting and retargeting return on ad spend targets should be different. For prospecting campaigns, as the user by definition has not visited your site, it’s pretty likely that the conversions you get from these campaigns are incremental–customers you wouldn’t have converted otherwise.

While it’s possible the customer had previously seen your ad on another channel and was eventually going to convert without the prospecting ad, most brands have ad spend concentrated on relatively few channels and so this is unlikely. Unless you’re running a ton of media across many channels, prospecting customers are very likely incremental.

…Not so for retargeting customers. If someone’s visited your site already, which they have to have in order to be retargeted, they are much more likely to buy your product anyway without seeing another ad. It’s not necessarily the case that they would have–some customers will have forgotten about your brand and the retargeting campaign actually causes them to convert.

But many customers don’t buy on their first site visit and are just waiting for the right time to buy when they need the product or have funds for it. Retargeting these customers, at best, slightly pulls forward their purchase.

What’s a good heuristic for what percentage of retargeting customers are actually incremental–implying how much lower your retargeting target should be? It depends on the brand but at Hubble it was about 45% lower. That was the point at which the incremental amount of spend got us a positive LTV customer net of marketing spend. But for your brand it could be 25%, or even 75%. It’s definitely not 0%, though.

This logic also applies, by the way, to other types of pseudo-retargeting on ad platforms. People who liked your brand’s Facebook page, or watched 75% of one of your videos also have the same incrementality issue. They’re probably not as likely to otherwise convert as a site visitor, but they are definitely less likely to convert than someone who hasn’t engaged with your brand at all. If you’re going after these buckets of people, it’s worth testing their incrementality too.

Don’t make the mistake of equalizing your ROAS targets across all of your campaigns. Figure out what percentage of spend for a campaign is incremental and then set the targets accordingly. Retargeting should be way lower–otherwise you’ll waste a ton of money.