Ask most multifamily finance teams how they arrive at next year’s rent number, and the answer involves lease expiration curves, renewal conversion assumptions, and concession burn-off modeled by unit group. Ask the same team how they arrive at next year’s ancillary revenue number, and the answer is often a flat percentage applied to the rent line, carried forward from last year with a small adjustment.
That gap is not a data problem so much as a treatment problem. Non-rent revenue gets budgeted like a bonus rather than modeled like a P&L line with its own drivers. This article lays out what a more rigorous forecast requires: which inputs matter, how to build the assumptions, and where the model needs to work differently at the property level than at the portfolio level.
Why ancillary revenue is difficult to forecast
Rent forecasting benefits from decades of shared methodology. Ancillary revenue forecasting does not have the same infrastructure behind it, which is part of why it gets treated as an afterthought.
The Revenue Method’s guide to multifamily forecast accuracy makes the underlying point directly: base rent alone does not tell the full revenue story, because ancillary income, amenity premiums, parking, storage, and fees can create <cite index=”87-1″>meaningful variance between projected and actual performance</cite>. That variance is not random. It comes from treating a revenue stream with its own volume, conversion, and pricing dynamics as if it moves in lockstep with rent.
It does not move in lockstep with rent, and the reasons are structural. Rent forecasting has a small number of well-understood drivers: units, lease terms, renewal rates, market rent growth. Ancillary revenue is driven by move volume, service adoption, and vendor pricing, three variables that behave differently from each other and from rent. A portfolio can hold rent growth flat while move volume climbs, or vice versa, and a forecast that treats non-rent revenue as a fixed percentage of rent will miss both directions.
There is also a discipline gap that is worth naming plainly. Freddie Mac’s 2025 Multifamily Outlook observed that when rent growth and occupancy trade off against each other, <cite index=”104-1″>operators have typically chosen the inverse, maintaining occupancy levels with less rent growth</cite>. That is a deliberate, modeled tradeoff, made because rent forecasting infrastructure exists to support it. Non-rent revenue rarely gets the same deliberate tradeoff analysis, not because it matters less, but because the modeling infrastructure to support that analysis usually is not there.
Historical vs event-based forecasting
Two different forecasting logics are available for ancillary revenue, and most operators default to only one of them.
Historical forecasting extends last year’s ancillary revenue forward, adjusted for a growth rate. It is simple to build and easy to explain, and it is also the method most exposed to the variance problem described above, because it assumes the relationship between revenue and its underlying drivers stayed constant. If move volume shifts, historical forecasting will not see it coming until the actuals already show the miss.
Event-based forecasting starts somewhere else: from the calendar of move-in and move-out events the portfolio actually expects, and builds ancillary revenue up from there. This is closer to how rent forecasting already works. Rentana’s guidance on budget-season revenue forecasting makes the same point about rent: a forecast built only on historical averages can understate risk, while mapping actual lease expiration timing into the model produces a materially more accurate quarterly pattern than smoothing vacancy into one annual percentage. The same logic applies directly to ancillary revenue. A portfolio with a large share of move-outs concentrated in a single quarter, for reasons tied to lease expiration timing that the rent forecast already accounts for, should expect ancillary revenue to follow that same concentration rather than a smooth monthly average.
The strongest forecasts combine both. Historical performance sets the baseline conversion and adoption rates; the event calendar determines when and how much volume those rates get applied to. Neither logic alone is sufficient, in the same way that neither is sufficient for rent.
Volume, conversion, and revenue-per-transaction inputs
A useful way to structure the ancillary revenue forecast is to borrow the three-input structure that sales and revenue operations teams have used for years to forecast pipeline: volume, conversion, and value per transaction.
The pipeline velocity formula used in B2B revenue forecasting multiplies the number of opportunities by the win rate and the average deal size, then divides by the length of the sales cycle, to <cite index=”91-1″>provide a real-time, data-driven revenue forecast</cite>. The specific formula belongs to a different industry, but the three-input structure underneath it, volume, conversion, and value, maps cleanly onto ancillary revenue.
Volume is the number of eligible move events, move-ins and move-outs, expected during the forecast period. This comes directly from the lease expiration schedule and renewal assumptions already built into the rent forecast, which is one reason ancillary revenue forecasting should sit next to rent forecasting rather than in a separate spreadsheet.
Conversion is the share of eligible residents who transact in a given service category, movers, packing, storage, insurance, or connectivity. Conversion should be tracked separately by category, because a resident who books a mover behaves differently from one who adds insurance verification, and blending the two into a single conversion rate obscures which category is actually driving or dragging performance.
Revenue per transaction is the average value of a completed transaction in each category. This is the input most exposed to vendor pricing changes and should be reviewed on a cadence independent of the broader forecast, since a pricing change from a single vendor can move this number without any change in resident behavior at all.
Multiplying these three inputs by category, then summing across categories, produces a forecast built from the same drivers that actually determine the result, rather than a single blended growth rate applied to last year’s total.
Property-level forecasting
A portfolio-level ancillary revenue forecast built from portfolio-level averages will systematically hide the properties that need attention most.
The logic here parallels a point made in utility budget forecasting, a different non-rent revenue category with its own volatility. Billee’s analysis of multifamily utility forecasting describes how portfolio-level data enables <cite index=”76-1″>cross-property benchmarking, identifying which properties are running above or below comparable properties</cite>, along with variance checks that flag properties whose growth rate deviates from the portfolio average. Ancillary revenue forecasting needs the same cross-property discipline. A portfolio average adoption rate of 30 percent could mean every property is converting at 30 percent, or it could mean half the portfolio is converting at 45 percent while the other half is converting at 15 percent. Those are very different operational stories, and only property-level forecasting distinguishes between them.
Property-level forecasting requires the volume, conversion, and revenue-per-transaction inputs described above to be tracked at the property level, not just the portfolio level. A property with an unusually high concentration of move-outs in a single month should show that in its forecast; a property with historically strong mover adoption but weak insurance adoption should carry different category-level conversion assumptions than a property with the opposite pattern. Rolling these differences up into a single portfolio number before the forecast is built erases the information a revenue leader would actually use to act.
Portfolio aggregation
Once property-level forecasts exist, aggregating them into a portfolio number is largely mechanical, but two decisions determine whether the aggregate is useful.
The first is whether the aggregation preserves category-level detail or collapses it. A portfolio forecast that reports a single ancillary revenue number tells a CFO less than one that reports movers, packing, storage, insurance, and connectivity separately, summed across properties. The category breakdown is what makes the forecast actionable, because different categories respond to different levers, pricing, vendor selection, resident communication timing, and a blended total obscures which lever to pull.
The second is scale. The NAA, IREM, and BOMA 2024 Income/Expense IQ benchmarking report covered financial data for <cite index=”21-1″>over 1 million multifamily units across more than 4,600 properties in 109 metropolitan markets</cite>, illustrating the range of property sizes and market conditions a single portfolio-level forecast is often asked to represent. A portfolio spanning that kind of range in submarket conditions, asset age, and resident demographics should not expect a single conversion rate or a single revenue-per-transaction figure to hold across every property. Aggregation should sum property-level forecasts built from property-level assumptions, not apply one assumption set across a portfolio and call the result an aggregate.
Forecast variance analysis
A forecast that is never checked against actuals is a projection, not a forecast. Rentana’s guidance on revenue forecasting is direct on this point: a single-point estimate can imply more certainty than the underlying assumptions support, which is why the strongest practice is to build a base case alongside upside and downside scenarios rather than anchoring to one number.
For ancillary revenue, variance analysis should run at the same three-input level used to build the forecast. If actual revenue misses the forecast, the first question is not whether the total was wrong, but which input was wrong: did fewer move events occur than expected, did conversion in a specific category come in below assumption, or did revenue per transaction shift because of a vendor pricing change. Each of those misses points to a different fix. A volume miss is a leasing and retention issue. A conversion miss is an adoption and resident-communication issue. A revenue-per-transaction miss is a vendor and pricing issue.
This is also where property-level detail earns its keep. A portfolio-level miss that looks modest in aggregate can be masking a large miss at a handful of properties offset by overperformance elsewhere. Reviewing variance at the property level, by category, on the same cadence used for rent variance review, is what turns a forecast into a management tool rather than a once-a-year budgeting exercise.
CFO dashboard
A CFO does not need every input in the model surfaced on a dashboard. Four numbers, reviewed on a consistent cadence, give a complete picture without requiring a finance team to rebuild the underlying model every time someone asks a question.
Forecast vs actual, by category. Movers, packing, storage, insurance, and connectivity, shown separately, not blended into a single ancillary revenue line. This is the single fastest way to see which category is driving a portfolio-level miss.
Variance by input. For any category showing meaningful variance, a breakdown of whether the miss came from volume, conversion, or revenue per transaction. This turns a variance number into a diagnosis rather than just a flag.
Property-level range. The spread between the highest- and lowest-performing properties on forecast accuracy, not just the portfolio average. A wide range signals that property-level assumptions need review even when the portfolio total looks close to plan.
Scenario bands. Base, upside, and downside cases for the current forecast period, so the dashboard shows a range rather than a single point estimate that implies more certainty than the assumptions support.
None of these require new source systems. They require the volume, conversion, and revenue-per-transaction inputs described above to already be tracked by category and by property, which is the same underlying requirement that makes the forecast itself possible.
Frequently asked questions
Can ancillary revenue really be forecast with the same rigor as rent?
Not with identical inputs, since ancillary revenue is driven by move volume, service adoption, and vendor pricing rather than lease terms and market rent growth. But the same discipline applies: build the forecast from its actual drivers rather than a flat percentage of rent, and review variance by input rather than only by total.
What is the biggest mistake operators make in ancillary revenue forecasting?
Treating it as a fixed percentage of the rent line rather than a revenue stream with its own volume, conversion, and pricing inputs. That approach hides which lever actually moved when the forecast misses, and it means the forecast has no mechanism for catching a shift in move volume or vendor pricing until the actuals already show the damage.
Should ancillary revenue be forecast at the property level or the portfolio level?
Both, but the property level comes first. A portfolio forecast built by aggregating property-level forecasts, each with its own volume, conversion, and revenue-per-transaction assumptions, preserves the operational detail that a single portfolio-wide assumption set erases.
How often should the forecast be reviewed against actuals?
On the same cadence used for rent variance review, tied to the same lease expiration and budget milestones. Ancillary revenue variance that goes unreviewed for a full budget cycle means a full year of misallocated attention before anyone catches which input actually moved.
Closing
Rent gets forecast with lease-level precision because the infrastructure to do that has existed for decades. Non-rent revenue can get the same treatment, but only if it is modeled from its actual drivers, move volume, service adoption by category, and revenue per transaction, rather than carried forward as a flat percentage of rent. Property-level forecasting, honest portfolio aggregation, and variance analysis that points to a specific input rather than a vague miss are what turn ancillary revenue from a budgeting afterthought into a forecastable P&L line.
If you manage 10,000 units or more and want to see what a category-level, property-level ancillary revenue forecast would look like for your portfolio, reach out to our team.
Related reading: how the move-in and move-out workflow produces revenue for property managers, and the complete guide to resident onboarding automation. See also Moved for multifamily operators and the resident experience.