Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Interactive Tools › ROAS Calculator

ROAS Calculator for Budget Cuts

Using the ROAS calculator for a budget cut: enter the revenue and spend of each line's last budget slice, then rank the slices against break-even ROAS.
Tutorialintermediate7 min readUpdated

Why a budget cut changes the calculator's inputs

The default page's inputs describe a whole period, so the ROAS that comes out is an average over every slice of that period's spend, and it usually reads higher than the return on the final slice. The marginal returns and forecasting guide gives the reason: in most channels results follow a curve of Diminishing Returns, so the final increment buys less than the ones before it.

A cut removes exactly that final increment. The cross-channel budget-cutting guide therefore ranks every line by what its last increment has been earning, and points out that a channel with a high average can sit near saturation while one with a modest average still has room.

The guide to reducing spend on Meta supplies the substitute: compare one line at two budget levels in periods alike in season and otherwise unchanged, then take the change in revenue over the change in spend. That slice is what the inputs should hold.

The inputs for a budget cut

Calculator inputOn the default pageFor budget cutsWhere to find it
Revenue from adsRevenue attributed to the ads in the periodThe extra revenue between two comparable budget levels for one line, from the business's own orders where possibleShop or CRM records for two periods of similar season, or a planned budget step
Ad spendAd spend for the periodThe extra spend between the same two levels: the slice a cut would removePlatform billing for both periods
Gross marginEveryday gross marginThe Contribution Margin of what each line sellsFinance's contribution margin for the products each line promotes

The revenue row changes most. Take it from the business's own orders wherever they can be tied to the line, because platform-credited revenue moves with attribution as well as with spend, and choose periods close enough in season that the calendar did not do the work: the marginal returns guide warns that Seasonality, promotions and competitor activity all move results independently of spend.

The margin row sets the bar. Build it as the contribution margin and break-even guide does, net of the costs that come with each sale; dividing one by that margin gives Break-even ROAS. Lines selling different products each get their own margin rather than a blend.

Entering the budget cut inputs step by step

  1. Agree what the cut protects and how much it must save

    The cross-channel guide first settles the objectives that survive, the products and customers the contribution depends on, and whether the saving is wanted as cash this quarter, as profit or as a smaller cost base. Write the answers and the amount down, and list every commitment separately, since the marketing budget cut checklist stops a cut that has no commitment list, and the cross-channel guide ends booked media and contracts at their break points or notice dates rather than breaking them early.

  2. Find a budget step for each variable line

    For every line bought by the click or the impression, find two recent periods at different budget levels, alike in season and offers, and record each period's spend from billing and revenue from the business's own orders.

  3. Enter each slice and note its return and profit

    Enter each line's difference in revenue and in spend with the contribution margin. The ROAS is the slice's marginal return, and the gross profit line beneath it, which reads as contribution after ad spend once contribution margin is entered, says what the slice earned or lost.

  4. Rank the slices and forecast the loss

    Order the slices from the one furthest below break-even ROAS and cut down the list until the saving is reached. Record the evidence, the forecast and what would reverse the decision in the budget reallocation memo.

A worked budget cut with illustrative inputs

Illustrative example — not a real result

Grivmoss Camp Kitchen ranks three slices before a cut

Grivmoss Camp Kitchen, an invented online seller of camping stoves and cookware, has to take $6,000 a month out of paid media for a quarter. The three lines under review sell the same range, so one contribution margin applies: at 40%, break-even ROAS sits at 2.5. Each variable line had a budget step in comparable months with no promotion running. Paid social prospecting went from $14,000 to $18,000 a month and its shop revenue from $61,000 to $68,200: a slice of $4,000 for $7,200. Category search went from $8,000 to $10,000 (revenue $36,600 to $41,000) and Shopping from $9,000 to $12,000 (revenue $39,600 to $49,200).

Entered one at a time, the slices ($4,000 for $7,200, $2,000 for $4,400 and $3,000 for $9,600) return 1.8, 2.2 and 3.2. The social slice loses $1,120 a month after ad spend ($7,200 × 40% − $4,000) and the search slice $240, while Shopping's earns $840, although social's average ROAS at $18,000 is a healthy-looking 3.79 and search's 4.1. Grivmoss removes the social and search slices, $6,000 in all. At the margin it forecasts $11,600 a month of revenue lost ($4,000 × 1.8 + $2,000 × 2.2), or $4,640 of contribution, so contribution after ad spend should rise by about $1,360 a month; on average returns the same cut would have seemed to lose about $23,360 of revenue a month. Untested brand search and retargeting stay at their current level, recorded in the memo as an assumption, and the search cut follows the social cut by a fortnight, each read on the shop's own orders.

Prefilled with the example's inputs: Revenue from ads $7,200 · Ad spend $4,000 · Gross margin 40% → ROAS 1.8× · Break-even ROAS 2.5× · Gross profit after ad spend -$1,120.

Interpreting marginal ROAS when cutting a budget

Read the ROAS output as the return on the slice and compare it with break-even ROAS, not with the line's reported figure or an account-wide target.

Gross profit after ad spend shows what each slice costs to keep. The budget cut checklist calls ranking lines on their average, instead of on their last increment, the commonest failure, and a negative profit line shows why: removing the slice adds its loss back to profit, so cut from the slice furthest below break-even and work down the list.

For the forecast, multiply the spend removed by the slice's ROAS. The Meta guide sets out this estimate; because the cut begins with the weakest spend, the volume given up for each dollar saved should grow as the cut goes deeper. Because the slices removed earn less than their lines' averages, this total sits below a forecast built on averages, and it is the figure to put in the cut memo. The Meta guide adds that a prospecting cut shrinks the retargeting pool it feeds, so part of the loss can arrive later in retargeting's results. When money returns, the marketing budget calculator page for scaling spend shows how a budget step is priced at its marginal cost rather than the average.

Treat every slice's return as a range. Because the calendar and competitors shift results independently of spend, a slice just below the bar is a closer call, while one far below it is a clearer case.

The calculator cannot see causation. Brand search and retargeting are often credited with sales that would have happened anyway, so the checklist keeps them at a level recorded as an assumption unless an Incrementality test exists; the ROAS calculator page for retargeting campaigns enters a holdout's incremental revenue once one does.

Finally, judge the cut against its forecast on the business's own orders once each staggered change has settled, as the Meta guide advises, since, on Meta, reported results can fall faster than real sales when retargeting's easy credit falls away. Booked media follows the media plan replanning tutorial, and the conditions for restoring spend go in the memo as a signpost and threshold from the scenario planning framework.

Pitfalls when cut numbers go into the calculator

  • Ranking lines on their average ROAS. The cut removes the last slice; the average blends it with every slice that came before it.
  • Reading a slice from unlike periods. A step taken across a sale, whose different margin the ROAS calculator page for discount promotions handles, or across a change of season measures the offer or the calendar, not the spend.
  • Trimming each line by an equal share. An even cut takes the same share from a line whose last slice still pays as from one whose last slice loses money; the Google Ads budget-cutting guide makes the same point inside one account.
  • Ranking brand search and retargeting on reported return. Without a holdout, their slices carry credit for sales that were coming anyway.
  • Changing every line at once. A slice's forecast can be checked against the business's own orders only when its cut lands alone, and on some platforms several learning periods would start together; space the changes out and read each once it has settled.

Frequently asked questions

Why not simply cut the line with the lowest ROAS?

Because a line's reported ROAS is an average over all of its spend, and a cut removes only the last slice. Saturation can leave a line that looks strong overall earning little on its final slice, while a line that looks ordinary may still have room to grow. Rank the slices rather than the lines and the cut gives up the least.

What if a line has never had a budget change to read?

Make one. Keep the line steady for a few weeks, then move its budget and hold the new level for a similar stretch in a comparable season; the difference between the two periods is the slice the calculator needs. Until then, signs of saturation, such as rising costs or climbing frequency in the same audience, hint at a poor margin, and an incrementality test of a budget step, if one exists, outranks them because it reads the marginal cost directly.

Should brand search and retargeting go first because their ROAS looks inflated?

Not on the reported figure in either direction. Attribution tends to credit both with orders that were coming regardless, which leaves their reported return a poor guide to the last slice. Without a holdout, choose a level for them, write it into the cut memo as an assumption, and schedule a holdout so the next decision rests on evidence rather than on credit.

Terms used on this page

Tools to try

Related