The Ecomm Analyst

Every brand I work with has, at some point, asked me to put a clean number on their influencer program. Most of them don’t like the answer they get.

The honest answer is that influencer attribution is the worst-attributed channel in ecommerce. The link gets clicked or it doesn’t. The discount code gets used or it doesn’t. Either way, you’re missing the much larger group of people who saw the post, didn’t click, didn’t use the code, and bought from you 19 days later through a Google search for your brand name.

If you measure influencer through last-click, you’ll undercount it by a factor of 3 to 5, depending on the category. If you measure it through code usage, you’ll undercount it by a similar factor and also create perverse incentives where your team starts pushing creators to be louder about the code, which destroys the content.

The approach I’ve settled on, which I run in ThoughtMetric, combines a few things and accepts that the answer is approximate.

The first input is direct attribution. Code use, link clicks, anything trackable. This goes into the bucket as a floor.

The second input is brand search lift in the 7 days after a post goes live. If a meaningful creator drops content on a Tuesday and branded search volume is up 40% on Wednesday and Thursday, that’s not coincidence. I’ll attribute a portion of that lift to the post, conservatively, knowing I can’t be precise about it.

The third input is new-customer mix during the post window. If the percentage of orders from new customers spikes after a creator post, especially from new customers who don’t have any other clean attribution, I’ll allocate some of those to the post too.

None of these are precise. Combined, they get you something better than last-click and worse than a randomized controlled trial.

The way I talk about influencer with clients now is to set expectations early. We will not have a clean ROAS number. We will have a directional view of what’s working and what isn’t. We’ll know within 60 days whether a partnership is moving the needle. We won’t know exactly how much.

This is uncomfortable for clients who came up in performance marketing, where everything is supposed to be measurable to the cent. The honest reality is that influencer is closer to a brand-building activity than a performance channel, and it should be budgeted that way. If you’re spending influencer money out of your performance budget and expecting performance returns, you’ll cut the program after three months and miss the slow compound it produces.

The creators who actually move the needle, in my experience, are the ones the brand has worked with for at least a year. The first post does almost nothing. The fifth post is when the audience starts to feel like the partnership is real. By the tenth post, you’re getting the lift the program was supposed to produce.

Patience is the actual unlock. Not measurement.