we had this unexpected win last quarter. partnered with a really sharp creator in mexico, campaign performed 40% above target. the brand loved it, the numbers were there, and now everyone wants to know: can we scale this?
the thing is, when i try to document what actually made it work, i keep running into the same wall. was it the creator? the timing? the specific product positioning? the fact that we built a real relationship instead of just transacting? probably all of it, but i can’t figure out which levers actually transfer to other creators, other products, other markets.
we tried to replicate it with a different creator in florida last month, and it just… didn’t click the same way. now i’m wondering if what we documented was the process or just a really good story i’m telling myself.
how are you guys capturing what actually works in one partnership so you can rebuild it somewhere else without just hoping lightning strikes twice? and at what point in the campaign cycle should you even start thinking about this?
the key mistake i see agencies make is trying to document the playbook after the campaign ends. you need to be building it in real-time. assign someone to document every single decision: why you picked that creator, what positioning angle you tested first, what feedback changed the narrative, which deliverables moved the needle.
think of it like archaeology—you’re not just collecting the finished artifacts, you’re tracking every assumption that got validated or invalidated along the way. that’s where the real playbook lives. in my shop, we actually built a simple scorecard for each partnership: creator profile match, positioning fit, content format, engagement mechanics, timeline. then we score it retrospectively and compare across projects.
the florida creator probably failed because you matched on one dimension (geography) but missed on others. that’s actually valuable data.
this is a scaling problem disguised as a documentation problem. before you can document anything, you need to isolate variables. was the mexico creator successful because of audience demographics, content style, your level of handholding, the product itself, or market conditions?
solid approach: run a controlled experiment. find a second mexico creator with similar audience profile but different style. replicate exact same product, positioning, timeline. if results are similar, you know the creator type matters more than individual personality. if results tank, you know it was about the person.
also—40% above target is great, but what was the baseline? if the baseline was weak, maybe you just got average results and got lucky with expectations. need that context to build anything repeatable. how are you tracking attribution?