“The greatest danger in times of turbulence is not the turbulence. It is to act with yesterday’s logic.” Peter Drucker
I have been doing this since 1999.
Toolbars. Coupon wars. Network consolidation. Mobile. The year influencers were going to replace the rest of us. Every one of those changed the job, and every one took years to do it, so you had time to see it coming, complain about it at a conference, and still adjust before it cost you anything.
This one is moving in months.
I am not going to write you another essay about how AI changes everything. You have read forty of those and not one of them told you what to do on Monday.
So here is the part nobody says plainly.
The work getting automated right now is not the work you are paid for.
Pulling lists. Hunting down a name and an email at a site you already knew you wanted. Checking whether a link still resolves. Writing the ninetieth version of the same outreach note. Rebuilding last month’s report. That is most of the week for most affiliate managers I know, and none of it is why anybody hired them.
At SmarterChaos I maintained a spreadsheet of publisher URLs by hand, on Sundays, and I want you to know I was proud of it.
The people who see this coming get a very good eighteen months. The ones still running the week the way they ran it in 2023 lose partners to somebody who is not smarter than them, just faster.
Not because they are lazy. Because they are waiting.
Waiting is the most expensive item on the menu right now.
What You’ll Get in This Issue
What to do this week with no budget and no new headcount. Why waiting on your network to build this for you is the costliest patience in the business. The honest answer on building your own with Claude or ChatGPT, and where it falls apart. Plus an invitation to watch all of it run live and unrehearsed on Thursday.
Part One: The Job Changed. The Job Description Did Not.
The affiliate manager role was built around a shortage of information.
Who are the publishers here. How do I reach them. What did they do last month. Are they still breathing. Every piece of that job existed because finding out was slow and manual, so we hired humans to find out.
That shortage is over. It ended quietly, sometime in the last two years, without a press release.
What is left is the part that was always the job. Deciding who is worth the money. Knowing which partner picks up the phone for you in December. Negotiating. Telling a brand the truth about their landing page. Choosing what not to do.
That work is not getting automated. It is getting more valuable, because the supply of everything else just went infinite.
The catch is the whole issue in one sentence. Reclaiming hours only helps if you move them somewhere better. Automate your busywork, refill the gap with more busywork, and congratulations, you have purchased a faster treadmill.
The gray hairs in this business, and gray hairs are experienced people, know the difference between a fad and a floor shift. A fad changes what you talk about at the conference. A floor shift changes who is still at the conference.
This is the second kind.
Part Two: The Playbook
Three moves. No budget approval, no committee, all three start this week.
Move 1: Audit the Week Before You Buy Anything.
Most people run this backwards. They ask “what AI should we get,” buy a subscription nobody opens, and produce a slide that says we are exploring AI.
Start with your calendar instead.
For one week, log where the hours go, in four buckets. Looking things up. Drafting and formatting. Reporting and admin. Talking to an actual human being.
You will not enjoy the ratio. Almost everybody I have watched work spends the week in the first three buckets and gets graded entirely on the fourth.
Now take the biggest bucket that is not human conversation, pick one repeatable task inside it, and run it with a tool for two weeks. Time it both ways. Keep it or kill it on the number, not the vibe.
Do it in that order, because it makes you literate. You cannot evaluate this stuff from a demo, where somebody else picked the example. The only way to find out where these tools lie to you with total confidence is to have used one badly yourself.
→ Do not buy a strategy. Automate one task you hate, measure it honestly, then go find the next one.
Move 2: Do Not Wait on the Platforms.
The question I get most is some version of “isn’t our network going to build this?”
Probably. Eventually. In a release note. And it still will not save you.
They build for the average of everybody, so what ships is useful to thousands of programs and decisive for none of them.
They can only show you what is inside the walls. A network surfaces publishers who are already in that network, and the partners who change your quarter are frequently the ones who are not. That boundary is a business decision, not a technical one.
And when it finally ships, your entire competitive set gets it the same morning. That is not an edge. That is table stakes with a press release.
None of which makes platforms the villain. Use them for what they are excellent at, which is tracking, settlement, payments, compliance, and disputes. That is real infrastructure and you should not want to build it.
Just stop filing your growth problem under somebody else’s roadmap.
→ Let the platform own the plumbing. Own the growth work yourself, and never let another company’s release schedule become your strategy.
Move 3: Build What Is Yours. Buy What Belongs to Everybody.
The other question, usually from the sharpest person in the room: why not just build it ourselves with Claude or ChatGPT?
You should. Go do it this week.
Build the assistant that drafts outreach in your voice instead of the voice of a press release. Build the thing that turns your weekly numbers into the summary your boss actually reads. That is an afternoon, it costs almost nothing, and the literacy you pick up is worth more than the automation.
Zero sarcasm, and yes, I am aware I sell software.
Here is where it stops working.
It is a demo, not a system. It runs beautifully on your machine, in your login, with you sitting right there to catch it when it wanders off. Add a second person or a week of vacation and you are quietly running an IT department you never applied to.
The model is not the hard part. The data is. A general model does not know which publisher actually drives revenue in your category. It knows what the internet says about them, and it will hand you the wrong answer in a beautifully written paragraph. That is the failure to be afraid of. Not the tool that breaks. The tool that is fluent, wrong, and never mentions which one it is being.
So here is the line, and it has held up every time I have used it.
Build what is specific to you. Your voice, your templates, your reporting quirks. Nobody else is ever going to build that.
Buy what is specific to the industry. Partner data, contact records, fit scoring, the judgment about which kinds of partners suit which kinds of brands. That is a collection problem measured in years, and you are not going to out gather it on the side while also running a program.
→ Build the part only you could build. Buy the part that took somebody else a decade to collect. Do not confuse the two, because one of them is fun and the other one is expensive.
Your Weekly Chaos Challenge
Two hours this week, neither of them costing anything.
Hour one, log your week into the four buckets and be honest about the ratio. Hour two, take the most repetitive thing you found and run it through a tool, badly, on purpose, until you know exactly where it starts making things up.
→ Automate one task you hate before Friday, then spend the hours you get back on a conversation you have been putting off.
And in the comments: what part of your week would you hand to a machine tomorrow with zero guilt?
YOU are invited THIS WEEK: Come Watch This Run Live
Thursday, September 10, noon eastern. Thirty minutes including questions.
No slides about the future of the industry. We take one real program from the registration list and our product team builds an actual recruiting campaign on camera. Partner matching and scoring, finding the right human at each partner, drafting the outreach, launching it. Unrehearsed, which is a polite way of saying you get to watch whatever happens.
Alfie’s job is a narrow one and we like it that way. Find and score partners, find the right people there, draft the outreach you approve. You keep every decision that matters, which is this whole issue built into a product.
Recording goes to everybody who registers, live or not.
Final Thought
One caution before you go automate your week.
Volume without judgment is just spam with better grammar. These tools make it trivially easy to point enormous output at real people, and the moment you do that without a human deciding it is worth their attention, you have used a great tool to become the thing everybody deletes. The tool is never the trick. What somebody does with it is.
This shift does not punish people for being slow. It punishes people for being certain.
The ones who come out of it in better shape will not have the biggest tool budget. They will be the ones who spent an afternoon getting their hands dirty, formed an opinion from something they actually did, and then moved their week around it.
Low bar. Also one most of your competitors will not clear this quarter, because it is easier to wait for somebody to announce the right answer.
Nobody is going to announce it. Nobody is coming to build it for you.
Go break something small this week.
Until next week.
Run toward the chaos.
Matt Frary
Chief of Chaos
President & COO, XPFlow
#AffiliateMarketing #PerformanceMarketing #PartnerMarketing #AI #ChaosToGrow
