How to Start an Affiliate Program Without an Agency (the Tool Is the Easy Part)

Starting an affiliate program in-house takes three things: tracking software (from about $49 a month, or free through Shopify Collabs), a commission set from your gross margin, and a recruited first cohort of partners. The software is the easy part. Most new programs stall because nobody recruits, and the ones that do grow often leak a discount code within weeks.

I think the "just hire an affiliate agency" advice skips a step. An agency makes sense once you know which partner types actually sell your product. Before that, you are paying someone a retainer to find out something you could learn yourself in about 90 days, with a spreadsheet and some slightly awkward emails. This piece is the in-house version: what to buy, what to pay, who to recruit first, and the two mistakes that quietly kill small programs.

If you are still deciding between a flat fee, a rev share, or a mix for creators specifically, start with our breakdown of creator compensation models. This article assumes you have decided that at least part of your partner program pays on performance.

The tracking software costs $49 a month. Recruiting is the actual job.

The tooling side has gotten cheap enough that it should not be a decision you agonize over. For SaaS and subscription businesses, entry plans for Rewardful start at $49 a month and Tolt at $69, and both plug into Stripe. For Shopify stores, Shopify Collabs is free to install, with a 2.9% processing fee on automatic commission payouts. That is basically the whole software bill for year one.

One small warning on Collabs. It sits at 4.1 stars across 426 reviews, but 21% of those reviews are one star, which is a pretty lopsided spread for a free first-party app. It works fine for gifting and basic commission tracking. Reviewers asking for tier systems and more detailed commission rules is a recurring theme, so plan on possibly migrating once you pass a few dozen active partners.

Networks are the other route. Awin, CJ, Rakuten, and impact.com give you a marketplace of publishers who can find you, plus fees on top of commission. Honestly, for a brand with no affiliate history, I would skip the network in the first quarter. The publishers most eager to join a brand-new program tend to be coupon and cashback sites, and (as the next sections get into) those are the partners you least want defining your program's economics.

Here is what the tool will not do: find anyone. A Shopify merchant posted in the Shopify Community after launching a Refersion program with commissions of up to 15% and a 90-day cookie window, asking how to get affiliates to actually sign up. Those are generous terms. The rate was not the problem. A program with a nice signup page and no outbound recruiting is just a form on a website.

Benchmark to aim for: 20 personally recruited partners in the first 60 days, with at least 5 of them producing a sale. If you hit 5 of 20, you have a channel worth tuning. If you hit 0 of 20, the problem is almost always the offer or the partner fit, and a bigger commission rarely fixes it.

Set the commission from margin, then launch below the ceiling

Most first-time programs pick a commission by looking at a competitor and adding two points. That works until you realize the competitor has a 70% gross margin and you have 35%.

Do the math the boring way first. Take your average order value, subtract cost of goods, shipping, payment fees, and expected returns. What is left is your contribution per order. Your maximum commission is whatever share of that you are willing to spend to acquire a customer, and it should be lower than what you already pay on paid social per new customer, otherwise the channel is not doing its job.

For reference, ReferralCandy's benchmarks put most DTC brands at 10 to 15% per sale, or a flat $10 to $15 on new-customer orders, with apparel at 8 to 15% and electronics lower at 5 to 10%. Digital products and software run much higher because the margin supports it.

Two specific moves I would make on day one:

• Pay more for new customers than returning ones. A partner who reminds an existing customer to reorder deserves something, but it probably should not be the same rate as a net-new buyer. If your tracking tool can distinguish them, use it.
• Launch below your ceiling. If the math says you can afford 15%, open at 10% and hold the other 5 points for a performance tier. Raising a rate is a nice email to send. Cutting one is how you lose your best partners (Amazon's affiliates found out how that feels when Amazon cut commissions up to 50%).

A quick worked example. A $70 average order with $38 in product, shipping, and fees leaves about $32 in contribution. A 12% commission costs $8.40, roughly a quarter of that margin. If your blended paid-social cost per new customer is sitting around $25 or higher, $8.40 looks very reasonable. If it is $6, the affiliate channel is the expensive one and you should know that before you scale it.

Codes leak. Links mostly don't.

This is the mistake I see recommended most often, weirdly, in "how to start" guides: give every affiliate a personal discount code and track sales by code. It is easy to set up. It is also the fastest way to end up paying commission on sales nobody referred.

Codes get scraped, posted to deal sites, and auto-applied by browser extensions at checkout. In a Shopify Community thread on coupon leakage in affiliate programs, merchants describe codes turning up on deal sites and Reddit, sending commission to the wrong partner and teaching customers to leave the checkout page to go hunt for a code. The consensus fix in that thread fits in four words: "No code, no leak." Bind the discount to the affiliate link so it applies automatically, and stop handing out strings of text that can be copied.

The scale of this is not small. Discount and coupon publishers took 42.4% of US affiliate revenue in the first half of 2025, according to Awin data cited by eMarketer, up from 39.7% a year earlier. Some of that is legitimate. A lot of it, in my opinion, is last-click credit for purchases that were already happening. We covered how messy this gets in the Honey and Rakuten standdown story, and that fight is between large, well-staffed companies. A new program with one part-time manager has far less ability to police it.

So the practical rule: link-based attribution by default, single-use or auto-applied discounts where you need an incentive, and no public coupon partners until you have at least three months of clean data on what your content partners drive on their own. You can always add coupon sites later. Removing them once they are 40% of your "affiliate revenue" is a much worse conversation with your CFO.

Your first 20 affiliates are probably already buying from you

Cold-recruiting affiliates is slow, and the reply rates are grim. The better first cohort is people who already like the product. In rough order of how well this tends to work:

1. Customers who already post about you. Search your brand name on Instagram, TikTok, and Reddit. Anyone who tagged you without being paid has already proven they will talk about the product. Email them first.
2. Small creators in your exact niche. Not the biggest ones. Our coverage of creator deal economics found 80% of creator deals pay under $300, which means a lot of mid-sized creators will seriously consider a well-structured commission plus product. A creator with 8,000 followers who reviews exactly your category can outsell one with 200,000 who reviews everything.
3. Newsletter writers and bloggers in adjacent categories. Evergreen content keeps sending clicks for months, which makes them worth the extra onboarding time.
4. Existing referral program users. If you already run a customer referral loop, your top referrers are affiliate candidates. We wrote about referral programs that actually scale, and the overlap between the two channels is bigger than most teams assume.

Send personal emails, not a mass blast. Include the commission, the cookie window, a product sample offer, and one sentence on why you picked them specifically. It sounds obvious. Most brand outreach still reads like a mail merge, and creators can tell.

For context on where this is heading: impact.com's 2025 State of Affiliate Marketing report, based on 818 marketers across eight countries, found 74% of brands earn between 11% and 30% of total revenue from affiliate partnerships, and 59% plan to put at least a quarter of their affiliate budget into creators. That is a survey of brands with established programs, so I would not expect those numbers in year one. But it does suggest the creator side is where recruiting effort pays off, more than the coupon side.

The paperwork that sinks programs around month six

This part is dull and I am going to keep it short, but skipping it causes real trouble later.

Disclosure. The FTC treats an affiliate commission as a material connection, and its Disclosures 101 guide for influencers is explicit that disclosure has to be clear and hard to miss. Brands can share responsibility when partners do not disclose, so put the requirement in your terms and actually spot-check posts. The FTC also keeps a longer Endorsement Guides FAQ if you want the full version.

Terms worth writing on day one:

• No bidding on your brand name in paid search (otherwise affiliates end up competing with your own ads for customers who were already searching for you)
• No coupon-site or browser-extension promotion without written approval
• A 30 to 60 day commission lock so returns and chargebacks are netted out before payout
• Your right to change rates with 30 days of notice

None of that is exciting. All of it is much easier to put in place before you have 50 partners than after.

FAQ

How much does it cost to start an affiliate program?

The software can be free (Shopify Collabs, plus a 2.9% payout fee) or roughly $49 to $69 a month for SaaS-focused tools like Rewardful or Tolt. The real cost is commission plus the time to recruit and manage partners, which is usually a few hours a week in the first quarter.

Should I use an affiliate network or run it in-house?

For a brand with no affiliate history, in-house first is usually the better learning path. Networks give you reach, but early reach on a network tends to skew toward coupon and cashback publishers. Move to a network once you know which partner types convert for you.

What commission rate should a new affiliate program offer?

For most physical-product DTC brands, 10 to 15% per sale, or a flat $10 to $15 per new customer, is the common range. Set yours from contribution margin, then launch a few points below your ceiling so you have room for a performance tier.

Where I would put the first 90 days

If I were starting one of these next week, I would spend maybe 10% of the effort on tooling and commission design, and the rest on recruiting and watching what the first 20 partners actually do. The first-quarter data tells you far more than any benchmark table, including the ones linked above. I would also keep the coupon sites out for a while, which will probably feel like leaving money on the table. Some of it probably is. I would still rather learn what my own partners can sell before paying anyone for customers who were already sitting on the checkout page.

By Notice Me Senpai Editorial