Programmatic advertising forecast assumptions

What to take away

A programmatic advertising forecast is useful when it makes the full cost base, chosen return measure and uncertain assumptions explicit.

The first mistake most teams make is quoting a return figure before writing down the full cost base, so nobody can check it later.
Build the forecast from media spend, working media ratio, conversion value and a margin you can defend, plus the click and lead rates that turn impressions into leads.
Label every assumption, including the ones you expect to be wrong, and date them.
If the forecast depends on a partner you have not vetted, the return figure is not yours to promise.

Set the cost base first

Start with everything the client pays, not just the media line: platform fees, agency time, creative production, tracking setup and data costs. A forecast built on media spend alone will overstate the return.

Cost lines to include

Platform fees
Agency time
Creative production
Tracking setup
Data costs
Mark unknown costs as unknown

Write each cost as a monthly figure in pounds and keep the list to five lines or fewer. Mark any cost you cannot estimate as unknown.

Choose the return measure

Return on investment needs one measure, not four. Pick the one the client already reports against, such as gross profit or booked revenue, and agree it in writing. For ROI, use attributable gross profit; if the agreed measure is revenue, apply its agreed margin first. Revenue divided by cost is not ROI.

Choosing the return measure

Can the client value a conversion?

Yes

Use the client's own measure, agreed in writing

Use a labelled illustrative figure, shown as an assumption

If the client cannot value a conversion, use a labelled illustrative figure, such as a £60 lead value, and show it as an assumption rather than a result.

State the window too. Mixing a 30-day return with a 12-month return is the most common source of dispute.

Build the worked example

Take a client with a monthly budget of £12,000 (illustrative). Assume 80 per cent reaches working media, so £9,600 buys impressions. At a £6 cost per thousand impressions, that is 1.6 million impressions.

Assume a 0.9 per cent click rate, giving 14,400 clicks. Assume 4 per cent of those become qualified leads, so 576 leads. At £60 per lead, that is £34,560 in attributed value. Suppose the client agrees in writing that half of that value is gross profit (an assumption), giving £17,280 of attributable gross profit.

Now add the full cost base from the first step. The £12,000 budget covers media and platform fees. Assume £1,500 for agency time, £800 for creative production, £300 for tracking set-up and £400 for data costs, so the full monthly cost is £15,000.

The attributed-value-to-cost ratio is £34,560 divided by £15,000, roughly 2.3 to 1, but that is not ROI. ROI is attributable gross profit minus full campaign cost, divided by full campaign cost: (£17,280 - £15,000) / £15,000, or about 15 per cent.

Test the assumption you trust least. If the lead rate is 3 per cent instead of 4, leads fall to 432, gross profit to £12,960 and ROI to about -14 per cent. A forecast should show both cases. Every figure above is illustrative, not a benchmark, and each should be replaced with the client's own data.

Check the audience assumptions

Audience size caps everything above. If the addressable audience is smaller than the impression volume implies, the click and lead rates will not hold.

ONS figures on home internet and social-media use can provide broad context for whether an audience definition is plausible. They do not show how much inventory a programmatic buying platform can reach, so use platform-side availability data for that forecast assumption, such as the reach estimates in Google Display & Video 360, The Trade Desk or Amazon DSP, whichever platform the plan uses.

Check the programmatic buying assumptions

State whether the plan uses an open auction or a programmatic direct deal, then label assumptions for auction CPM, available inventory and any supply-path costs; verify these against the buying platform and contract.

Vet the supply chain before you sign

A forecast is only as good as the inventory behind it. Check who sells the impressions and whether they appear on a regulator's list of problem advertisers.

The Advertising Standards Authority maintains a non-compliant online advertisers list that is worth checking during due diligence on partners and networks.

If a partner fails that check, remove them from the plan and rebuild the forecast.

Review at day 14 and day 30

Set two review points before launch. At day 14, compare actual cost per click and cost per lead with the forecast. At day 30, decide whether to hold, shift budget or stop.

Keep the original forecast visible. A forecast quietly edited to match results teaches the team nothing and makes the next pitch harder.

Common questions

How do I calculate programmatic advertising return on investment?

Subtract the full cost, including fees and production, from the attributable gross profit, then divide by the same full cost. Dividing value by cost gives a return ratio, not ROI. State the attribution window and conversion value too, because the result means nothing without them.

What figure should I use for working media ratio?

Use the ratio in the client's current contracts if one exists. Otherwise pick a conservative illustrative figure, label it as an assumption, and replace it with the contracted ratio once the paperwork is available.

How often should the forecast be updated?

Update it at day 14 and day 30, then monthly. Record what changed and why, so the client can see whether performance or a revised assumption moved it.

Can I promise a return before the campaign runs?

No. Present the figure as a forecast with named assumptions. Promising a return before you have verified the inventory, audience and conversion value creates a commitment you cannot support.