An affiliate program has two funnels and most teams only watch the first one. The application funnel ends with an approval, and an approval feels like a win: the list grew, the dashboard counter moved, the program looks healthy. The second funnel ends with a link in public and a click attached to it, and that one is usually empty.
The gap between the two is not enthusiasm. It is instructions. A new affiliate has a to-do list that existed before you emailed them, and your program is item twelve on it. Onboarding is the work of moving it to item one, or accepting that it will sit at twelve forever.
TL;DR
- Approval is a permission, not a partnership. The affiliate cannot promote what they cannot describe, and most silence is missing information rather than missing motivation.
- Measure activation, not applications. Activated means one public placement plus one tracked click inside 30 days, counted by approval cohort.
- The sequence is four touches: day 0, day 2, day 7, day 30. Anything denser reads as pressure and gets archived.
- A kit is a link, approved claims and two visuals. A folder of logos is not a kit, and it is the most common thing sent in place of one.
- Ask for one specific piece of content with a brief. "Write about us" produces nothing, "compare us to the tool you already pay for" produces a draft.
Approval is not activation
An approval email grants a tracked link and a rate. It does not grant context: what the product is for, who it is not for, what claim is safe to repeat, or where this particular person's audience actually lives. The affiliate has to assemble all of that before they can write a sentence, and assembling it is work they were not asked to do.
That is why the first message matters more than the rate. A generous commission on a product the affiliate cannot explain in one line produces nothing. A modest commission with a clear one-line description and a ready-made comparison produces a post.
There is also a sequencing trap. Teams approve in batches because reviewing applications is tedious, then send one generic welcome to forty people, then measure the result thirty days later and conclude the channel is weak. The batch is the problem. Forty approvals in a day is forty onboarding conversations collapsed into a broadcast, and a broadcast cannot answer the question each affiliate is privately asking: what exactly do you want me to do.
The one metric for month one
Revenue is the wrong number for a new affiliate cohort. Sales cycles are longer than the onboarding window, commissions arrive late, and a single refund can make a small cohort look broken. The number that tells you whether onboarding works is activation.
Activation rate = affiliates with at least one tracked click in their first 30 days / affiliates approved in the same week.
Two details make it useful. Cohort by approval week, not by calendar month, so a slow week does not contaminate a fast one. And count clicks, not posts: you cannot see every placement, but a tracked click is evidence that a placement exists somewhere.
Track time to first click alongside it. A cohort that activates in four days on average is being onboarded. A cohort that activates on day 28 is being reminded, which is a different and more expensive process.
Inside Sled, the dashboard records share kit activity: copies and opens of the pre-written posts, per affiliate. That gives you a leading signal before the first click. An affiliate who copied a post twice and produced no click has published somewhere you cannot see, probably a private community or a DM, or pasted a link without the code. Both are fixable in one message. An affiliate with no copy and no click has not started, and that is a different message.
Day 0: the welcome that sets expectations
One email, five facts, one question. Not a PDF, not a video tour, not a link to a help center with fourteen articles.
| Element | What it says | Why it belongs |
|---|---|---|
| The link | Your tracked URL, ready to paste | The first thing anyone tries |
| The rate | Percent and whether it recurs | The question they will otherwise ask |
| The window | Cookie or attribution length | Determines what they can promise an audience |
| The payout | Minimum and available methods | Decides whether the program is worth their time |
| The ask | One sentence: where do you plan to share this | Turns a broadcast into a conversation |
The question at the end is doing most of the work. "Where do you plan to share this?" is easy to answer, it commits nothing, and the answer tells you which brief to send next. Someone who says "my newsletter, 2,000 subscribers, mostly indie founders" needs a different second email than someone who says "my deals site".
Send it from a person, not from the program. Replies to a program address feel like replies to a form. Replies to a founder get read. The welcome email template has the full copy, including the parts that set expectations about how often you will contact them, which is the thing affiliates actually resent when it goes unstated.
Day 2: a kit, not a folder
Most kits are a zip file with fourteen logo variants and a brand color document. That is a design archive. A working kit answers the three questions a person needs before they publish: what do I paste, what am I allowed to claim, and what image fits the place I am posting.
- A link they can paste, per placement if possible. One link is fine. A link buried in a table inside a PDF is not.
- Three approved claims, written as sentences they can copy verbatim. "Syncs with Polar, Stripe and Paddle" is a claim. "The best affiliate platform" is a liability you just handed to someone else.
- Two visuals in the right ratios, one square, one wide. Not the full brand pack.
- One disclosure line, ready to paste, matching the platform's rules and yours.
Sled's share kit exists for exactly this reason: pre-written posts per platform with placeholders for the link, program name and commission, so a new affiliate can publish something accurate in two minutes instead of drafting from scratch. The design principle behind it is in the share kit guide, and the short version is that the first post has to be easier than not posting.
Do not ask for approval on every post. Affiliates read approval requirements as a delay, and delay is what kills week one. Give them claims that are true and a disclosure line that is compliant, then let them publish.
Day 7: send a brief, not a topic
"We would love a blog post about us" is not an assignment. It has no audience, no angle, no length and no promise, so the affiliate has to invent the whole thing and will postpone it until the invention feels worth doing.
A brief is five lines:
- Audience. Who is reading, in their words. "Founders who just launched on Polar and are paying attention to their first ten customers."
- Angle. The specific comparison or problem, not the product. "Why a flat fee beats a percentage of commission at low volume" is an angle. "Our product" is not.
- Must mention. Three facts, each verifiable on your site today.
- Avoid. The one claim that is technically true but invites an argument, the competitor you do not want named, the price you do not want quoted out of context.
- Link and code. With the disclosure line repeated, because it will be pasted.
One brief, one piece. Not a content calendar. The affiliates worth keeping have their own calendar, and the fastest way to lose them is to look like a second job.
The type of piece should follow the affiliate, which is where most programs generalize too early. A reviewer writes a comparison. A newsletter operator writes a paragraph in an issue that already exists. A customer advocate writes a short post about the problem they had. A deals site wants a code and a deadline, not a brief. Sending a review brief to a deals site is how you get silence with a read receipt.
Day 14: the check-in that is not a nag
By day 14, the silent affiliates have split into two groups that look identical in the dashboard and need opposite treatment.
The first group is stuck. They tried the login, hit the payout form, saw a tax field they did not understand, or could not find the link again after closing the tab. Their silence is friction, and it is your bug.
The second group is deprioritizing. They meant to write, the week got away, and every reminder makes the task feel heavier.
The message that works for both is a question about the work, not a request for status. "Did the payout setup go through, or did it ask for something unclear?" is a question with a fix attached. "Just checking in, any update?" is a request for a confession, and it gets no reply.
Before sending anything, read the signals you have. Share events, clicks, whether they opened the welcome email, whether they ever set a payout method. An affiliate who has not configured payouts has not started, and no amount of encouragement fixes a form they have not reached. The right move there is to remove the form from the critical path: payouts can be configured after the first commission, and the first message should say so.
Day 30: activate, park, or part ways
At day 30, sort the cohort into three buckets and act on each once.
Activated. At least one click. Send a short note naming the specific thing they did, and one next step that is bigger but still concrete: a second placement, a comparison page, a newsletter slot. This is the moment most programs go quiet, right when the affiliate has proved they can execute and would accept a second assignment.
Warming. Signed in, copied a post, no click. One diagnostic question, then one fix. If there is still nothing at day 45, move them to parked.
Dormant. No sign-in, no share event, no click. Stop the sequence. Do not remove them: removal in Sled soft-deletes the affiliate, blocks the same email from applying again and kills the link, which is the right tool for fraud and the wrong tool for indifference. Park them, keep the door open, and revisit when something changes that is worth an email: a new integration, a price change, a feature their audience asked for.
The discipline here is that dormant affiliates cost you attention every time you open the dashboard. A program with 300 approved and 12 active is not a large program, it is a small program with a misleading header. Parking is how you get the header to tell the truth without burning the relationship.
Three affiliate types, three onboarding paths
The sequence above is the default. It is wrong for at least two of the five types worth recruiting, and the mismatch usually shows up as "this affiliate ghosted us" when the real issue is that they were sent the wrong first ask.
| Type | What they need in week one | First ask | Signal they are alive |
|---|---|---|---|
| Writer or reviewer | Product access, three claims, an angle | A comparison against a tool they already use | A draft outline or a question about the product |
| Newsletter operator | A short blurb, a code, a send date | One paragraph in an upcoming issue | A reply with the issue date |
| Customer advocate | A reason to tell the story, a simple link | A short post about the problem they solved | A share event on day one or two |
| Deals or coupon site | A code, a discount, an expiry | Listing the offer with terms | A listing URL and a click spike |
| Developer or educator | Sandbox, sample data, a scripted demo | A tutorial built on the sample data | A repo, a gist, or a question about the API |
The pattern: writers need an angle, operators need a slot, advocates need permission, deals sites need an offer, developers need a working environment. None of those are the same email.
If you are still choosing which types to recruit, the field guide to affiliate types covers the tradeoffs, and the honest starting point for most small programs is customers and writers, because both already understand the problem you solve.
Technical products need a shorter path, not a longer brief
The failure mode unique to developer tools and infrastructure is asking for a review of something that takes an afternoon to install. The affiliate is willing, the setup is not, and the draft dies in a terminal.
Fix the path instead of the copy:
- Give a sandbox with data already in it. A demo account that shows a real dashboard beats a free trial that starts empty.
- Record the three-minute version. Not a product tour. The exact path from signup to the thing that makes someone say "oh".
- Write the demo script as a numbered list the affiliate can follow while recording.
- Offer to do the first setup call. Twenty minutes with you removes the entire installation objection.
For a product like Sled, the equivalent is a program with real conversions in it: an affiliate explaining attribution is much more convincing when their own dashboard shows a click, a conversion, and a commission, even if the commission is on a test order. Affiliate programs for developer tools goes deeper on the recruitment side of this.
Six ways onboarding fails quietly
None of these produce an error message. All of them produce a program that looks fine and pays nobody.
- Bulk approval with a broadcast welcome. Forty approvals, one email, zero conversations. The rate of reply is the tell.
- A kit without a paste-ready link. The affiliate has to hunt for the one thing they need, and stops.
- "Write about us" with no brief. The task expands until it is abandoned.
- Status requests instead of diagnostic questions. Each one makes the task heavier.
- Removing dormant affiliates to clean the dashboard. You lose the ability to re-engage them later, and you taught them the program is not serious.
- Measuring revenue in week two. Attribution takes longer than that, and the conclusion you draw is wrong before the data exists.
The fix for all six is the same shape: fewer messages, each one carrying something the affiliate can use, sent by a person who will read the reply.
FAQ
How long should affiliate onboarding take? Four touches over thirty days is enough for most products. Day zero for the link and the ask, day two for the kit, day seven for the brief, day thirty for the decision. If the product needs installation or a sandbox, add a setup call between day two and day seven rather than adding more emails.
Should I remove affiliates who never promote? No. Park them. Removal is for fraud, self-referral and terms violations, because it blocks the email from applying again and kills the link. Indifference is reversible, and a dormant affiliate who returns after a product change is a free activation. See why affiliates stop promoting for the patterns behind the silence.
Do coupon and deals sites need the same onboarding? No, and this is the mismatch that wastes the most effort. They need an offer with terms, a code, an expiry and a listing, not a content brief. Send them the commercial details on day zero and skip the kit entirely. Affiliate coupon leakage covers the risk side of that relationship.
What is a good activation rate for an affiliate program? Any number copied from someone else's program is noise, because approval standards differ wildly. What matters is your own trend: cohort by approval week, watch whether the curve is moving, and investigate the week it drops. A drop after a process change (new form, new kit, new payout requirement) is the most useful signal in the whole dashboard.
The first sale is not the goal of onboarding. The first click is. Get a new affiliate to publish one accurate link in their first week, and the sale becomes a matter of their audience, not your follow-up.
If you are rewriting the first email right now, the welcome template has the exact copy, and Sled gives you ten days to run this sequence on a real program.