If I were in your shoes, I'd immediately make an open-text "How did you hear about us?" question mandatory in the post-purchase survey.
That's where I'd start because discount codes only give you one piece of the puzzle. You're completely blind to all the "dark social" influence and word-of-mouth recommendations that don't involve a specific code. Michael Brandt made this point on The Bottom Line, explaining that a post-purchase survey is just as vital as direct code tracking for getting a real sense of your ROI. It's the simplest way to begin capturing data on sales you're influencing but can't directly attribute through a checkout code alone.
In week one, I'd focus entirely on getting the infrastructure right. This means creating unique, trackable discount codes for every single influencer, no exceptions. At the same time, I'd get that survey live. The crucial part is making the question required but open-ended, not a dropdown menu. You want the customer's raw, unfiltered answer. This is how you'll catch misspelled influencer names or even discover new affiliates you didn't know were driving traffic. I'd start a simple spreadsheet to manually tag survey responses to specific influencers. It's a bit of grunt work, but it’s the foundation for everything else.
By month one, with data flowing in, I'd start analyzing the overlap and the gaps. Look at which customers used a code and mentioned the influencer in the survey, that’s your most confident attribution. Then, more importantly, I'd zero in on customers who mentioned an influencer but didn't use their code. This is your incremental lift, the true halo effect. As Yash Chavan discussed on eCommerce Fastlane, this is how you start to measure the actual incrementality of your program. I'd also look for clues about discount code leakage, a challenge Connor Rolain highlighted on Marketing Operators, by spotting codes used by people who attribute their discovery to something else like a Google search.
Throughout this process, I would completely ignore the temptation to find a single, "perfect" attribution software that promises to solve everything. It's a distraction. Tony Chopp from Ecommerce Playbook is right that chasing perfect attribution is futile. No single tool can see the whole picture, especially with the complex halo effect of influencer marketing. Your blended system of codes and surveys, while a little messy, gives you a much more honest view. Don't get paralyzed trying to assign a precise value to every touchpoint.
The biggest trap to avoid is paying out commissions on sales that were going to happen anyway. If an influencer's code is constantly being used by repeat customers or is easily found on coupon sites, they aren't really acquiring new customers for you. The survey is your best defense against this. When a customer uses an influencer's code but credits "a podcast ad" or "a friend" in the survey, you know that influencer didn’t truly drive an incremental sale. Being disciplined about analyzing this lets you weed out ineffective spending and reinvest in true influencer partnerships that grow your business.