Why this matters
Every platform reports its own ReachUnique people who saw an ad., and each figure is credible only inside its own walls. Added together they describe more people than exist, and the people counted twice are the ones most likely to be over-exposed. The cross-platform reach and frequency guide explains why no figure can be deduplicated from outside; at this level the skill is planning with an estimate whose assumption is written down, and deciding what money should not be spent once the audience has seen enough.
The brand and last year's flight
Pomewick Cider is an invented regional cider brand sold in supermarkets and pubs. Its audience is adults of legal drinking age who drink cider or beer, and all creative goes through alcohol advertising rules before it runs, including age targeting on every platform.
Pomewick Cider: last year's six-week flight
Audience: 5,000,000 adult cider and beer drinkers in the brand's sales regions.
| Platform | Reported reach | Average frequency | CPM | Viewability |
|---|---|---|---|---|
| YouTube, bought direct | 1,900,000 | 4.2 | £9.00 | 88% of measurable |
| Meta video | 2,300,000 | 6.8 | £4.50 | Not reported on the same basis |
| TikTok | 1,100,000 | 3.5 | £5.50 | Not reported on the same basis |
| Open-web display | 3,400,000 | 2.9 | £1.60 | 70% measurable; 52% of measurable viewable |
The agency's wrap-up claimed 8,700,000 people reached. No lift study ran. A twice-yearly online tracker of 600 category drinkers showed unaided awareness at 11% before and 12% after.
This year's budget, draft plan and inventory options
Budget £240,000 for eight weeks before summer. The marketing manager's effective frequency range is 3 to 6 exposures per person over the flight.
| Line | Budget | CPM | Impressions | Planner reach | Planner frequency |
|---|---|---|---|---|---|
| DSP: YouTube and curated video deals | £130,000 | £13.00 | 10,000,000 | 2,500,000 | 4.0 |
| Meta video | £70,000 | £5.00 | 14,000,000 | 2,000,000 | 7.0 |
| TikTok | £40,000 | £8.00 | 5,000,000 | 1,250,000 | 4.0 |
- Meta's planner at £40,000: 8,000,000 impressions, reach 1,700,000.
- Open-exchange video through the DSP costs £7.00 per thousand; comparable inventory last year was 70% measurable and 52% viewable of measurable. Curated deals run at about 95% measurable and 85% viewable.
From platform reach to deduplicated reach
Buying YouTube through the DSP lets one system estimate reach across it and the curated deals; Meta and TikTok stay outside. For those, a modelled overlap is the honest option: assume how the audiences duplicate, state the assumption, and replace it with panel data when budgets allow. The simplest assumption treats each platform as reaching people independently. Heavy video users appear on every platform, so real overlap is usually higher, which makes the independent estimate an optimistic ceiling. The video awareness guide adds the rule that matters once the estimate exists: when the audience has reached the planned FrequencyAverage times a person saw an ad., extra money belongs somewhere else.
Task
- Show why last year's wrap-up figure cannot be people, and what it can still tell you.
- Estimate the draft plan's deduplicated reach and average frequency, and compare them with the range.
- Rebalance the plan and set the reach and frequency goal, with caps by platform.
- Choose the video formats and lengths for each line and the branding rule for every cut.
- Set the viewability and attention rules, comparing the open exchange and the curated deals on a viewable basis.
- Say how reach and brand lift will be read after the flight.
Worked answer
Reach, frequency and viewable cost arithmetic
- Last year: 8,700,000 claimed against an audience of 5,000,000. The sum is impressions grouped by platform.
- Draft plan: summed reach 5,750,000, again more than the audience. Independent overlap: unreached share 0.50 × 0.60 × 0.75 = 0.225, so reach is about 77.5%, or 3,875,000. Average frequency 29,000,000 ÷ 3,875,000 ≈ 7.5, above the range.
- Rebalanced: Meta at £40,000. Unreached 0.50 × 0.66 × 0.75 ≈ 0.25, reach about 75%, or 3,760,000. Impressions 23,000,000, average frequency about 6.1. Spend £210,000; £30,000 held back.
- Cost per thousand impressions known to be viewable: open exchange £7.00 ÷ (0.70 × 0.52) ≈ £19.20; curated £13.00 ÷ (0.95 × 0.85) ≈ £16.10.
Reading last year. Platform reach figures cannot be summed, and the display figure counts browsers rather than people. Each figure still compares formats within its own platform, and Meta's high frequency shows where the over-exposure sat.
Goal. Plan a deduplicated reach of roughly two thirds to three quarters of the audience, with average frequency inside the range. Even the rebalanced plan sits at the top of it, because the independence assumption is generous, so cap each platform separately below the range's top and set a campaign-level cap in the DSP. Hold back the remaining money for a lighter midsummer flight rather than piling more impressions into one.
Formats. Short non-skippable cuts on YouTube for repetition, a longer skippable cut for the story, and vertical sound-on cuts for Meta and TikTok made from the same shoot. The brand name, the bottle and the orchard colours appear in the opening seconds of every cut. Open-web display banners leave the plan: they carry no sound or motion and added reach nobody could verify.
Quality rules. Buy only measurable inventory, exclude in-app placements without a measurement SDK, use a pre-bid viewability threshold, and block made-for-advertising sites. The cheaper CPMCost per 1,000 impressions. on the open exchange is dearer per impression known to be viewable, so the curated deals win on cost as well as quality. Report time-in-view as measured and any attention score as modelled, compared only within one vendor, as the viewability and attention guide advises.
Reading results. Report the DSP's unique reach, Meta's and TikTok's separately, and the deduplicated estimate as a range with its assumption. Read the frequency distribution, not the average. Run a platform lift study on the largest line, run the tracker wave before and after the flight with enough respondents, and watch branded search on a lag. Judge creative only if reach landed as planned, as the media objectives and reach goals guide sets out.
Common mistakes
- Summing platform reach. It always exceeds the truth.
- Treating the independence estimate as fact. It is a ceiling; write it down as an assumption.
- Choosing inventory on CPM. Compare cost per viewable impression and attention.
- Spending past the frequency ceiling. Hold money back or extend presence instead.
Test your judgement in the reach and attention checkpoint, or revisit awareness planning at the intermediate brand marketing level.
