Affiliate marketing tips

How to Forecast Recurring Affiliate Income by Modelling Churn and New Referrals

Build a twelve-row spreadsheet in agency counts and see where your recurring SendTo commission levels off, how long that takes, and which lever moves the ceiling.

A calculator resting on printed spreadsheet pages with a pencil, lit by window light on a desk
Photo by Jakub Żerdzicki

Recurring commission grows only while new referrals arrive faster than old ones cancel. Every recurring book eventually levels off at a ceiling set by two numbers, your new referrals per month divided by your monthly churn. Three new agencies a month at 4% churn settles at 75 active agencies. It never goes past that, however long you wait, unless one of those two numbers changes.

If you've tried to forecast recurring affiliate income by taking this month's referrals and multiplying by twelve, you've overstated your first year and hidden the plateau completely. The fix is a twelve-row spreadsheet that counts agencies, not currency. It shows you where your SendTo commission flattens out, how many months it takes to get there, and whether more referrals or stickier referrals move the ceiling further.

We're SendTo, and we're using our own affiliate programme as the worked example because we know every term. This is written in the week of 10 August 2026, with the autumn peak in email marketing interest still ahead of you. Each step gives you the action, what you should see, and the mistake to avoid.

By the end you will have a twelve-month model that shows where your recurring commission levels off

The finished thing is one spreadsheet tab. It has a row per month and columns for starting agencies, new agencies, cancelled agencies, ending agencies and commission. A single input cell holds the partner subscription, and two more hold your referral rate and churn rate. Change any of those three cells and the whole forecast redraws itself. That's the point. You'll use it to forecast recurring affiliate income honestly, then to decide where your promotion time goes next.

Be clear about what the sheet is. It shows the shape of a recurring book under assumptions you choose. It's not a promise of earnings, and the figures in this article are example inputs, not results anyone should expect. Here's what to have open before you start.

  • A spreadsheet. Google Sheets or Excel both work, and every formula here runs in either.
  • Your own referral history from the SendTo affiliate dashboard, or a cautious guess if you're just starting.
  • The current partner subscription, taken from the pricing page. You'll need it only in Step 6.
  • Twenty minutes. The first version is quick, and the scenarios come later.

Step 1: Count in agency-months first, because prices change and the shape of your book does not

Open a new tab and type five column headers across row 1. Month, Starting agencies, New, Cancelled, Ending agencies. Put month labels down column A for twelve rows, starting with August 2026. Leave money out entirely until Step 6.

Why hold back the money? The partner subscription is set on our pricing page, and prices change over time. The compounding behaviour of your book doesn't. Three agencies a month at 4% churn bends into the same curve whatever the fee is. Model the agency counts first and you're modelling the part you control, which is who you refer and how many. The fee multiplies at the end. If you want the reasoning behind why a flat partner fee makes this kind of forecast tidier than usage-based payouts, we covered it in flat fee versus usage based programmes.

What you should see now is an empty grid, twelve rows deep, five columns wide. Nothing calculated yet. Above the grid, label three input cells for later, Partner subscription, New per month and Monthly churn.

Step 2: Set a realistic new-referral rate from your clicks, not your hopes

Start with last quarter. Pull your clicks and the number of agencies that became paying partners from the affiliate dashboard, then divide agencies by three to get a monthly rate. If you had 600 clicks and 6 paying agencies across May, June and July, your rate is 2 a month. Use that, not the 5 you'd like.

Only a few months of data, or none? Pick a cautious whole number, such as 1 or 2 new agencies a month, and label it as your guess. A small honest number teaches you more than a large hopeful one, because you'll replace it with actuals within a quarter anyway.

Timing matters as much as volume here. SendTo's attribution lasts 90 days from the first click, and the first touch wins. An agency owner who clicks your link in May, reads three more reviews, books a demo in June and starts paying in July still counts as your referral. So referrals lag your promotion by weeks or months. A busy month of content shows up in the New column one or two months later, not the same month. Put your referral estimates in the month agencies start paying, not the month you published.

Add a seasonal multiplier instead of a straight line

Interest in email marketing peaks in January, November and December, and drops lowest in June. A flat 2 a month ignores that. Add a row under your months for a multiplier, set 1.0 as a normal month, and multiply your base rate by it. These are example multipliers, and yours should come from your own traffic.

Example seasonal multipliers, averaging 1.0 across the year. Label them as your own assumptions.
MonthAugSepOctNovDecJanFebMarAprMayJun
Your multiplier0.81.01.01.21.21.31.11.01.00.90.7

Notice these average exactly 1.0 across twelve months. That's deliberate. Seasonality shifts when your agencies arrive. It doesn't move the ceiling, because the ceiling depends on your average referral rate over the year. With the 90-day window in mind, consider nudging the peak one month later, since November clicks often turn into December and January partners.

Step 3: Pick a monthly churn rate and apply it to the agencies you already have

Churn is the share of your active agencies that cancel each month. In the Cancelled column, the formula is Starting agencies multiplied by your churn cell. Apply it to Starting, not Ending, because an agency that joins this month hasn't had a chance to leave yet.

Always work in monthly churn. Annual figures hide how fast small monthly differences compound. At 2% a month, about 22% of a starting group is gone after a year. At 4% it's about 39%. At 6% it's about 52%. Two percentage points a month sounds trivial, and it's the difference between keeping most of a year's referrals and losing half of them. We've written separately about how software retention affects your income, so here we'll stay on the number you type in.

Which rate should you assume for agencies on a white-label platform? We won't hand you a statistic. Test 2%, 4% and 6% and see how much the answer moves. Then think about why partner agencies tend to stay. An agency on SendTo sells email marketing to its own clients under its own brand. Those clients log into the agency's branded portal, get billed through the agency's own pricing, and read their reports there. Leaving means moving every client's lists, forms, sending domains and reporting history, and explaining the switch to each of them. That's a real cost, and it's why 4% is a fair middle case rather than an optimistic one.

Step 4: Roll the months forward and watch growth slow on its own

Now the sheet comes alive. Two formulas do all the work, and you copy them down twelve rows.

Cancelled = Starting agencies x Monthly churn
Ending agencies = Starting agencies + New - Cancelled
Next month's Starting agencies = this month's Ending agencies

Here's the result with 3 new agencies a month, 4% churn, and nobody active at the start. Agency counts only, rounded to one decimal place for display.

Example inputs of 3 new agencies a month and 4% monthly churn. The sheet carries unrounded values, so a displayed row can differ by 0.1.
MonthStartingNewCancelledEnding
10.030.03.0
23.030.15.9
35.930.28.6
48.630.311.3
511.330.513.8
613.830.616.3
716.330.718.6
818.630.720.9
920.930.823.1
1023.130.925.1
1125.131.027.1
1227.131.129.0

Look at the gap between twelve months of referrals and what's left. You referred 36 agencies. You end the year with 29. The multiply-by-twelve forecast said 36, and it'll keep saying 72, then 108, while your real book does something very different.

Extend the formulas to 36 rows and the curve bends hard. Month 24 ends at about 46.8 active agencies. Month 36 ends at about 57.7. In month 1 your book grew by 3. By month 12 it grows by 1.9, and by month 36 by about 0.7. You're still getting 3 referrals a month, and growth still slows, because cancellations scale with the size of the book while new referrals don't. That's why recurring commission stops growing even when referrals keep coming.

Step 5: Find your ceiling with one division: new referrals ÷ churn

This is the number the whole sheet has been building towards. Divide your new agencies per month by your monthly churn and you get the level your book settles at. With the example inputs, 3 ÷ 0.04 = 75 active agencies. Put that formula in a cell beside your inputs and label it Ceiling. It's the fastest way to forecast recurring affiliate income over the long run, because it doesn't need a single row of the grid.

Why does it work? At 75 agencies, 4% churn cancels exactly 3 a month, and you add exactly 3 a month. Inflow equals outflow, so the book stops growing. Below 75, you add more than you lose. Above it, you'd lose more than you add. The book drifts towards the ceiling from either side and stays there while your inputs hold.

Getting there takes longer than you'd think. Starting from zero at 4% churn, you reach half the ceiling, about 37.5 agencies, after roughly 17 months. You reach 90% of it, about 67.5, after roughly 57 months. That answers the question of how long a recurring book takes to level off. Years, not months, and the last stretch is the slowest.

Doubling referrals and halving churn give the same ceiling at different speeds

You've got two levers. Double referrals to 6 a month at 4% churn and the ceiling doubles to 150. Halve churn to 2% at 3 a month and the ceiling also doubles to 150. Same destination. Very different journeys.

Example inputs only, in agency counts. Month 12 figures start from zero active agencies.
ScenarioCeilingActive at month 12Months to half the ceiling
3 new, 4% churn7529.0about 17
6 new, 4% churn15058.1about 17
3 new, 2% churn15032.3about 34

More referrals show up in your book straight away. Lower churn shows up slowly, because it only protects agencies you've already got, and early on you don't have many. Over five years the two cases converge. Over one year, referrals win by a distance. So is it better to get more referrals or to refer agencies that stay longer? Early, more referrals. Later, stickier ones. Step 8 turns that into a rule. The reasoning behind targeting agencies with retainer models is the long-run half of this table.

Step 6: Convert agency-months into commission with SendTo's terms, exactly as they stand

Now, and only now, money goes in. Open the pricing page, read the current partner subscription, and type it into your single Partner subscription input cell. Add a Commission column to the right of Ending agencies and point every row at that one cell.

Commission = Ending agencies × Partner subscription × 20%
Example (row 2, fee in B1):  =E2*$B$1*0.2
Ceiling commission:          =Ceiling*$B$1*0.2

The 20% comes straight from our terms, and nothing else gets added to it. You earn 20% of the platform fee, the partner subscription, that each agency you refer pays SendTo, every month that agency pays it. You never earn a share of what that agency's own clients pay the agency. An agency can charge its clients whatever it likes for email marketing under its own brand, and that's their revenue, not part of your commission. Model the partner subscription only.

Timing needs one more adjustment. Commission is paid monthly, by PayPal. An agency that starts paying in October earns you commission for October, but the cash reaches your PayPal account in a later monthly payout. Add a Cash column that copies the Commission figure from the row above, so each month's earnings land one row later. Treat that one-month shift as your own assumption, and correct it once you've seen your first payout arrive. Combined with the attribution lag from Step 2, a click in August can become cash in your account months later. That's normal, and the sheet should show it.

What you should see is a Commission column that has exactly the same curve as Ending agencies, scaled by one number. If the shape changes, a formula is broken.

Step 7: Run three scenarios so the forecast survives a bad quarter

One forecast is a guess. Three forecasts are a range, and a range survives contact with real months. Right-click your tab, duplicate it twice, and name the copies Cautious, Expected and Strong. Change only the referral and churn inputs on each. The partner subscription stays the same across all three.

Example scenario inputs, in agency counts. Replace them with your own assumptions.
ScenarioNew per monthMonthly churnActive at month 12Ceiling
Cautious26%17.533.3
Expected34%29.075
Strong43%40.8133.3

Now find out which input matters most. Go back to the Expected tab and move one input at a time. Raise referrals from 3 to 4 at 4% churn and month 12 lands at about 38.7 agencies, with a ceiling of 100. Instead cut churn from 4% to 3% at 3 referrals and month 12 lands at about 30.6, also with a ceiling of 100. Both moves raise the ceiling by the same amount. Inside year one, the referral change does about five times as much work. Your first-year result is most sensitive to referral rate, and your long-run result is equally sensitive to both.

Seasonality belongs here too. Run your seasonal multipliers through all three tabs and look at June. A cautious June with a 0.7 multiplier adds very little, and that's a quiet month, not a broken forecast. If you want to plan your promotion around those dips, we've covered how to keep earnings steady through seasonal shifts.

Check it worked: three months of actuals should land inside your scenario range

A forecast earns trust by meeting reality. Add two columns to the Expected tab, Actual new and Actual cancelled, and at the start of each month paste in the figures from your SendTo affiliate dashboard. Add a third column for actual active agencies. After three months, compare that line against your Cautious and Strong tabs.

Inside the range means keep going. Don't tune the inputs to chase one good or bad month. Three months of data with only a handful of agencies is noisy, and one agency more or less moves the line a lot. If you only have a few months of data, this is how your forecast gets better. You replace guesses with actuals, row by row, and the inputs follow.

Outside the range means something specific is off, and you should find which number before changing anything.

  • Referrals below Cautious: check the lag first. Attribution runs 90 days from the first click and the first touch wins, so promotion from last month often hasn't converted yet. Compare clicks with your earlier quarters before cutting your referral input.
  • Churn above Cautious: look at who you referred. Agencies that signed up with no clients to onboard have nothing holding them, so change how you set expectations about what SendTo is for, which is running email marketing for an agency's clients under its own brand.
  • Commission not matching active agencies × fee × 20%: check that your fee cell matches the current pricing page, that every row points at that one cell, and that your Cash column carries the one-month payout shift.

Use the model to decide where your next hour of promotion goes

The sheet's real job is to tell you what to do next. Look at your current active agencies and divide by the ceiling cell. Far below the ceiling, under about half, new referrals win. Each extra agency you bring in shows up in next month's book, and churn hasn't had much to work on yet. Spend the hour on content and outreach that reaches new agency owners.

Close to the ceiling, the maths flips. Extra referrals still help, but each one partly replaces an agency you're losing. Lowering churn protects a large book at once. Spend the hour on referrals that stay, meaning agencies with retainer clients they'll run on the branded portal. Write content that sets honest expectations before the click, too. This is also where the 80/20 pattern turns up in affiliate work. A few pages and channels tend to bring in most of your referrals, and your Actual new column will show you which ones once you tag them. If you want your readers to see this maths themselves, the same formulas sit behind interactive ROI calculators.

It's mid-August 2026, and the autumn and January peaks in email marketing interest are still ahead of you. Build the sheet this week. Set your inputs honestly, label them as your own, and remember what the model shows. It shows the shape of a recurring book under your assumptions, not a promise of what you'll earn. Then point September's promotion at whichever lever your ceiling cell says matters most.

Affiliates earn 20% of the partner subscription each referred agency pays SendTo, every month it pays, with 90-day first-touch attribution and monthly PayPal payouts. Sign up and your referral link goes straight into the New column.

Join the SendTo affiliate programme

Frequently asked questions

How do I forecast recurring affiliate income with only a few months of data?

Take the paying agencies from your last quarter and divide by three. If you don't have that, pick a cautious whole number like 1 or 2 new agencies a month and label it as a guess. You'll swap it for real actuals within a quarter anyway.

What monthly churn rate should I plug in for agencies on a white-label platform?

Run 2%, 4% and 6% and see how far the answer moves. 4% is a fair middle case, because leaving SendTo means moving every client's lists, forms, domains and reports. If your referrals include agencies with no clients lined up, set churn higher.

Why has my recurring commission stopped growing when referrals keep coming in?

Your cancellations grow with the size of your book, but your new referrals don't. At 3 a month and 4% churn, your book grows by 3 in month 1, by 1.9 in month 12 and by about 0.7 in month 36. It levels off at 75 agencies.

How long does a recurring commission book take to level off?

Longer than you'd expect. At 3 new agencies a month and 4% churn, the ceiling is 75, but you're only at about 58 active agencies after 36 months. The ceiling is new referrals divided by monthly churn, and you approach it slowly from below.

Is it better to get more referrals or to refer agencies that stay longer?

Both double your ceiling. Going from 3 to 6 a month or from 4% to 2% churn lifts it from 75 to 150 agencies. Doubling referrals gets you there faster, because churn sets how quickly the book closes in on its ceiling.

How does the 90-day first-touch window change when referrals show up?

An agency that first clicks your link and starts paying within 90 days is still yours, even if it read other reviews in between. So referrals lag your promotion by a month or two. Log each one in the month the agency starts paying, not the month you published.

Should seasonality change my forecast?

It changes when your agencies arrive, not where you level off. Interest peaks in January, November and December and bottoms out in June, so use monthly multipliers that average 1.0. You'll also want to push the peak a month later to allow for the attribution lag.

Do I earn anything on what an agency charges its own clients?

No. You earn 20% of the partner subscription each agency you refer pays SendTo, every month it pays. You never get a share of what its clients pay the agency. Commission is paid monthly by PayPal. Join the SendTo affiliate programme at /affiliate-programme.

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