Affiliate marketing tips

How to Forecast Recurring Affiliate Income from Agency Referrals

Build a 36-month spreadsheet forecast of recurring agency referral commission in units, then test whether more referrals or better retention moves your total more.

Row of coin stacks getting taller from left to right on a wooden table in soft side light
Photo by PublicDomainPictures

Three numbers decide your recurring commission. They're how many new agencies you refer each month, how many of them are still paying a month later, and the 20% share of each agency's partner subscription. Put those three into a spreadsheet and you can forecast recurring affiliate income for any month you like, without guessing and without waiting a year to find out.

The model also tells you something you can't see from a dashboard. Adding referrals moves your total more for roughly the first three years. Retention sets a ceiling you can never climb past, however hard you push. So the input worth working on depends on how far ahead you're planning. A one-year plan and a five-year plan call for different work.

This guide walks you through building that forecast, step by step, using SendTo's affiliate programme as the worked example because it's ours and we know every term. Every figure below is a worked example to show the mechanics. None of them is a prediction of what you'll earn.

By the end you will have a 36-month forecast and know which lever to pull

You'll finish with one spreadsheet. The main tab runs 36 rows, one per month, showing active referred agencies, monthly commission and cumulative commission. A second tab, the sensitivity tab, runs the same model three ways side by side, so you can see whether one extra referral a month or two extra points of retention does more for you over your planning horizon.

It takes about half an hour. Before you start, gather these.

  • A spreadsheet. Google Sheets or Excel both work, and nothing here needs macros.
  • Your last 90 days of affiliate clicks, signups and paid conversions from your affiliate dashboard. If you're new, a careful best guess is fine, and Step 2 shows you how to make one.
  • The current partner subscription from the pricing page. You'll only use it at the very end, for reasons Step 1 explains.

Step 1: Count in commission units so the forecast survives price changes

Start by defining one commission unit. One unit is 20% of one month's partner subscription for one referred agency. That's it. If an agency pays SendTo for a month, you earn one unit for that month.

This works because of how the programme pays. You earn 20% of the platform fee, the partner subscription, that each agency you refer pays SendTo, every month that agency keeps paying. You never earn a share of what that agency's own clients pay the agency. An agency that resells SendTo to forty clients at its own prices earns you exactly the same as one with four clients. So the client count, the agency's markup and its usage don't belong in your model at all.

That leaves one thing to count, agency-months. Every paying agency in a given month is one unit. To forecast recurring affiliate income in units, you only need to know how many agencies are paying in each month.

What you should see is a column headed "units" where you might have expected a currency column. Leave it that way until Step 8.

If you want to go further and show prospects what the platform is worth to them, that's a different spreadsheet. Our guide to interactive ROI calculators for SaaS affiliate campaigns covers that side.

Step 2: Estimate new referrals a month from the funnel you actually have

Your first input is new paying agencies per month. If you already have a few months of referrals, count them and take the average. If you don't, build it from your funnel.

The funnel version multiplies three numbers. Clicks a month, times the share of clicks that sign up, times the share of signups that become paying agencies. As a worked example, 300 clicks a month at a 5% signup rate gives 15 signups, and 20% of those becoming paying partners gives 3 new agencies a month. Use your own rates from the last 90 days, not ours.

Lag the inputs to match the 90-day attribution window

Attribution lasts 90 days from the first click, and the first touch wins. That has two effects on your forecast.

First, conversions trail clicks. An agency owner who clicks your link in March, runs a trial with one client in April and starts paying in May still counts as your referral. So don't credit March with agencies it hasn't produced yet. If your typical gap from click to paid is six weeks, shift the new-referrals column down by one or two months from the click data.

Second, first touch means first touch. If someone clicked another affiliate's link before yours, that referral belongs to them, even if your review closed the deal. Your real conversion rate already reflects that, which is one more reason to use your own numbers.

What you should see is a single, realistic number, such as 3 a month. One number is enough for now. Step 5 bends it for the seasons.

Step 3: Set a monthly retention rate, and be stingy with it

Retention is the share of referred agencies paying this month that are still paying next month. In the worked example we use 95% as the base case and 97% as the better case. Pick one rate for your base, and add a pessimistic case a few points lower, such as 92%.

Why might agencies that resell SendTo stay a while? Because their clients run on it. An agency selling email marketing under its own brand to retainer clients has those clients' lists, sequences, landing pages and reports all living in its portal. Leaving means migrating every client, re-warming sending domains and explaining the change to each one. That's real friction, and it's why we dig into why agency retainer models drive higher affiliate LTV in a separate piece.

Friction isn't loyalty, though. Agencies close, merge or lose the clients they bought the platform for. We're not going to give you an industry benchmark to copy, because a benchmark from a different kind of product and a different kind of buyer tells you nothing about your referrals. If you have no data, use 95% as a working assumption, run 92% next to it, and replace both with your real figure after six months.

Monthly and annual retention are very different numbers

95% sounds high. Compounded over a year, it isn't. 0.95 to the power of 12 is about 0.54, so at 95% monthly retention only around 54% of a given month's referrals are still paying twelve months later. At 97% monthly, about 69% are.

For more on how churn behaves across software products and what it does to commission over time, read our explainer on how software retention impacts your income.

Step 4: Build the month-by-month table with one formula

The whole model is one line of arithmetic. Active agencies this month equal active agencies last month, times retention, plus new referrals this month. Units this month equal active agencies, since one paying agency is one unit. Cumulative units are a running sum.

active_t = active_(t-1) × r + new_t
steady state = new ÷ (1 − r)

Sheet layout (retention in H1, e.g. 0.95)
A: Month   B: New   C: Active   D: Units   E: Cumulative

Row 2 (month 1):  C2 = B2          D2 = C2   E2 = D2
Row 3 onwards:    C3 = C2*$H$1+B3  D3 = C3   E3 = E2+D3
Fill rows 3 to 37 down for 36 months.

Put 3 in every cell of column B and 0.95 in H1. Fill down. What you should see is a curve that climbs quickly and then flattens. Using the example inputs, you'll have about 27.6 active agencies at month 12, 42.5 at month 24 and 50.5 at month 36. Cumulative units after the first twelve months come to about 196.

The flattening is the steady state, new divided by one minus retention. At 3 a month and 95%, that's 3 ÷ 0.05, which is 60 active agencies. Each month you add 3 and lose 5% of everyone you have, and at 60 those two numbers cancel out. You approach 60 and never reach it. That ceiling is the most useful single number in the spreadsheet, because it tells you where your current inputs lead if nothing changes.

The common mistake here is compounding retention onto the new month's agencies, typing something like (C2+B3)*$H$1. That churns agencies before they've paid their first month and quietly lowers every row. New referrals are added after retention is applied, not before. The other mistake is hard-typing values into column C instead of formulas. That breaks the sensitivity tab you build next, because changing H1 then changes nothing.

Step 5: Bend the referral line for seasonality

A flat 3 a month is tidy, and the year doesn't work that way. Public interest in email marketing peaks in January, November and December and drops to its lowest point in June. Agency owners plan their year in January and build client campaigns for the run-up to Christmas, so that's when they go looking for a platform. Your new referrals follow the same shape.

Add a column of monthly multipliers next to your new-referrals column, then multiply the base figure by each one. Here's a starting set. Treat it as your assumption, not data, and replace it with your own pattern once you have a year of referrals.

  • January 1.3, February 1.1, March 1.0
  • April 0.95, May 0.9, June 0.7
  • July 0.8, August 0.85, September 1.0
  • October 1.05, November 1.15, December 1.2

Those twelve add up to exactly 12. That matters. Your multipliers should average 1, so the seasonal version moves referrals around the year without quietly adding or removing any. If yours add up to 13, you've given yourself a free extra month of referrals.

Now apply the lag from Step 2. Interest peaks in January, but with a six-week gap from click to paying agency, the January clicks turn into February and March agencies. Shift the multipliers by the same one or two months you used for the click data.

This is written in the week of 9 March 2026. That puts the spring and summer dip ahead of you, with June the low point, and the November to January lift as the stretch where new referrals cluster. What you should see is a lumpier new-referrals column and an active line that still rises, but in steps, flattening through summer and climbing again in winter. A slow July is in the model. It isn't a sign the forecast is broken. Our piece on retaining affiliate earnings during seasonal market shifts covers what to publish through the quiet months.

Step 6: Run the sensitivity test: one more referral or two points of retention?

This is the tab that tells you what to work on. Copy your main table three times on a new sheet. Leave the base case at 3 new agencies a month and 95% retention. Change one input in each copy. The second gets 4 a month at 95%. The third keeps 3 a month and moves retention to 97%. Use flat referrals here, without the seasonal multipliers, so the comparison stays clean.

Active referred agencies, each one commission unit a month, under three worked-example scenarios
ScenarioActive at month 12Active at month 24Active at month 36Ceiling
Base: 3 a month, 95%27.642.550.560
More referrals: 4 a month, 95%36.856.667.480
Better retention: 3 a month, 97%30.651.966.6100

Read across the rows. At month 12, the extra referral gives you 36.8 active agencies against 30.6 for better retention. At month 24 it's still ahead, 56.6 against 51.9. By month 36 the gap is less than one agency. The two lines cross at around month 38, and after that the better-retention case pulls away towards a ceiling of 100, while the extra-referral case flattens towards 80.

So the answer depends on your horizon. On a one or two-year plan, volume wins. Another referral a month shows up in your active count straight away, while retention gains compound slowly. On a plan longer than about three years, retention decides how high you can ever get. That's the whole point of building the forecast. If you forecast recurring affiliate income over twelve months and spend those months chasing stickier agencies, you've optimised for a horizon you aren't planning for. The reverse is just as true.

Where the 80/20 rule fits

The 80/20 rule in affiliate marketing says a small share of your sources, pages or partners produces most of your results. It's a rule of thumb, not a law, and the split is rarely exactly 80/20. In this model it shows up in two places. A handful of sources usually bring in most of your referrals, and a handful of referral types usually account for most of your long-staying agencies. The sensitivity tab tells you which of the two to chase, and Step 7 shows you how to find them.

Step 7: Weight referral sources by how long their agencies stay

One retention rate for every referral is a fair start. It hides the most useful thing in your data. Split your new-referrals column by source, such as your newsletter, consulting calls, community posts and search content, and give each source its own row in the model with its own retention rate. Total the active counts at the bottom.

If you don't have data yet, keep every source at your base rate and wait. After six months, look at which sources' agencies are still paying. Two patterns are worth watching for. Agencies that already sell retainers have clients ready to move onto the platform, so it becomes part of a service they're already paid for. Agencies that watched the product work before they signed up, on a call or in a demo, arrive knowing what they're buying. Referrals from freelance consulting calls often fall into both groups.

What makes an agency stay is what its clients see each month. SendTo's client portal reports real human opens, with security scanner opens excluded, alongside clicks and bookings. So the agency's clients see numbers they trust. Segments built from clickers, people who opened but didn't click, and non-openers give the agency a reason to send the next campaign. New sending domains get warmed up automatically, and volume is spread across several brand-matched domains. That means an agency's first client launch doesn't sink into spam and send them looking elsewhere. When you write about those things, you attract agencies that plan to run clients on the platform, and those are the ones that stay.

Faceless affiliate marketing forecasts the same way

Yes, you can do faceless affiliate marketing, with no video, no personal brand and no calls, and still forecast it. The model only needs clicks, conversion rates and retention. A comparison page or a written teardown is just another source row. Expect it to convert at its own rate, and give it its own retention figure once the data comes in, because the agencies it brings in arrive having seen less of the product.

Step 8: Turn units into money and payment dates

Only now does currency come in. Open the pricing page, take the current partner subscription and multiply it by 0.2. That's the value of one commission unit. Put it in a single cell, say H2, and add a money column next to units that multiplies each row by $H$2. If the price changes, you update one cell and the whole forecast follows.

Commission is paid monthly, by PayPal. Add a payment-month column next to each earning month. As a working assumption, set it one month later, then match it to the date your first PayPal payment actually lands. What you should see is a cash line running slightly behind the units line. That gap matters if you're budgeting against it, because a strong December pays you in a later month.

Working backwards from a daily target

The "hundred a day" question is better answered as arithmetic than as a promise. Take your daily figure and multiply it by 30.4 to get a monthly target. Divide that by the value of one unit. The result is the number of active agencies you need. Then multiply that number by one minus your retention rate. That gives the new referrals a month you'd need just to hold it.

Say your target works out at 40 active agencies. At 95% retention, holding 40 takes 2 new referrals a month, because you lose 5% of 40 each month. But 2 a month gives you a ceiling of exactly 40, which you approach and never reach. At 3 a month the ceiling is 60, and you'd pass 40 at around month 21. So set your referral rate so that your ceiling sits well above your target, roughly one and a half times it, or the target stays out of reach.

Check the forecast against reality every month, and fix the input that drifted

The forecast earns its keep once real numbers arrive. At the start of each month, add two columns next to the forecast, actual new agencies and actual active agencies, taken from your affiliate dashboard. Then compare each one with its forecast.

If new referrals miss, the problem is upstream. Go back to Step 2 and check clicks, signup rate and paid conversion separately to find which one dropped. Check the seasonal multiplier for that month before you panic, too. A June miss against a flat forecast is often no miss at all.

If new referrals hit target but the active count sags, retention is the problem. Look at which source rows are losing agencies and lower their rates in the model. Then decide whether to keep sending traffic from that source or move effort to the ones that stay. One bad month is noise. Three months drifting the same way is a signal, so re-set the input then and not before.

So what's the success rate of affiliate marketing? There's no single figure worth quoting. Any number covers different programmes, products, payout models and definitions of success, and it tells you nothing about your own funnel. Your forecast, checked against actuals for six months, gives you a better answer. It shows whether your inputs are moving towards the ceiling you want, and which one to change if they aren't. For the programme terms in full, read our review of the SendTo affiliate programme.

Your next step is simple. Build the 36-row sheet, run the three scenarios, and pick the lever that matches your horizon. You earn 20% of the partner subscription of every agency you refer, every month that agency pays SendTo.

Get your affiliate link and start filling in the real numbers.

Join the SendTo affiliate programme

Frequently asked questions

Does the 20% commission apply to what the agency's clients pay, or only to the platform fee?

Only the platform fee. You earn 20% of the partner subscription each referred agency pays SendTo, every month it pays. An agency with forty clients earns you the same as one with four, so client counts and markups don't belong in your model.

How long does a referral count for, and what if someone clicked another affiliate's link first?

Attribution lasts 90 days from the first click, and first touch wins. If someone clicked another affiliate's link before yours, that agency is theirs, even if your review closed the deal. Lag your new-referral column a month or two behind your clicks.

Should I chase more referrals or agencies that stay longer?

It depends on how far ahead you're planning. Over roughly the first three years, extra referrals move your total more. Retention sets the ceiling, though: at 3 new agencies a month, 95% retention caps you near 60 active agencies, and 97% lifts that to 100.

What retention rate should I assume if I have no data yet?

Use 95% a month as a working assumption and run 92% next to it. Replace both with your real figure after six months. Don't plug an annual rate into a monthly model, because 95% monthly leaves only about 54% of a month's referrals still paying a year later.

What is the 80/20 rule in affiliate marketing, and how does it fit a forecast?

In this model, it's a reminder that a few of your referral sources tend to bring the agencies that stay longest. Step 7 gives each source its own retention rate instead of one blended average, so you can see which channels earn your time.

Can I do faceless affiliate marketing and still forecast income?

Yes. The model only needs clicks, signup rate, paid conversion rate and retention. Reviews, comparison pages and newsletters all feed the same funnel, so a faceless channel forecasts exactly the same way.

How do I make $100 a day from affiliate marketing with agency referrals?

Work backwards. Multiply your daily target by about 30, then divide by one unit, which is 20% of the current partner subscription on the pricing page. That's how many paying agencies you need. Then check whether your ceiling, new referrals divided by one minus retention, ever reaches it.

When and how is commission paid?

SendTo pays commission monthly, by PayPal. Step 8 turns your monthly units into money and payment dates, so you can see when each month's commission actually lands in your account.

Earn 20% of what every partner you refer pays, every month.

Refer agencies, studios, consultants and resellers to SendTo. Your commission is 20% of the platform fee each partner pays, for as long as they pay it.

90-day attribution · Paid monthly by PayPal