Video Campaign Groups and Reach Planner in 2026: Planning YouTube Reach and Frequency
If you run more than one YouTube campaign at a time, you have had a problem that Google Ads gave you no way to solve until 2026: frequency was counted and capped per campaign. Three campaigns each capped at “3 per week” could deliver nine exposures to the same person, and you usually found out from complaints about repetition and quietly decaying performance. Video campaign groups in Google Ads close exactly that gap.
On 13 July 2026, video campaign groups with reach and frequency optimisation became globally available. Combined with Reach Planner, that gives you a plan → buy → verify loop across a whole pool of campaigns rather than one campaign at a time. This guide covers how groups work, which campaign types are actually steered by them, what frequency to target, and where the planner systematically misleads.
What video campaign groups are
A video campaign group ties several YouTube campaigns to a shared reach or frequency goal. Budgets and creative stay separate — you are not pooling money and you do not lose control of what runs where. Only the objective becomes shared.
Two objectives are available:
- Increase campaign group reach — the system works to maximise unique, deduplicated reach across the campaigns, trimming wasteful repeat exposure.
- Set a frequency target — the system steers toward an average of roughly 2 to 7 exposures per person per week.
The operative word is deduplicated. Until now each campaign counted reach inside its own boundary and audience overlap was invisible. Now the overlap is both visible and manageable.
Not every campaign type is steered
This is the detail most teams miss on the first setup. Campaign types fall into two tiers.
| Tier | Campaign types | What it means |
|---|---|---|
| Fully coordinated | Efficient Reach, Non-skippable Reach, Audio Reach | The system actively steers buying toward the group goal |
| Countable only | Video Views, Demand Gen video, Target Frequency, reservation buys | Impressions count toward the totals but are not steered |
Practical implication: if most of your volume sits in Video Views or in video inside Demand Gen campaigns, a group gives you accurate combined frequency counting but very little control. Control appears only when the group contains a meaningful share of reach campaigns the system can actually move.
What frequency to target
Google points to its own marketing mix modelling study across roughly 600 US brands using 2023–2025 data, which identified about 2.7 weekly exposures as optimal and associated it with a 19% ROI lift. Treat those figures as a vendor-stated benchmark rather than a standard: the study has not been independently verified, the sample is US-only, and the product itself offers a much wider 2–7 band.
A sensible working approach:
- Low (2–3) — broad reach objectives, brand introduction, large audience, long flight. The goal is maximum unique people.
- Mid (3–5) — product launches, seasonal pushes, narrower audiences, or a message complex enough to need repetition.
- High (5–7) — short flights with a hard deadline (a sale, an event), narrow audiences already close to action. Irritation risk rises, and that should be a conscious trade.
Test the frequency hypothesis with measurement rather than instinct: Brand Lift and Search Lift on YouTube show what happens to awareness and search interest at different exposure levels. Click metrics cannot answer that question at all.
Reach Planner: what it does and where it misleads
Reach Planner is a free media planning tool inside Google Ads. You set geography, audience, budget, formats and a period, and it returns a forecast.
What the forecast covers
- Impressions and unique reach;
- average frequency;
- TRPs (target rating points) and cost per point, for anyone planning in broadcast terms;
- conversion-based metrics for most campaign types;
- co-viewing impressions on connected TV — an important correction, since more than one person is often in front of that screen.
Limits worth knowing before you plan
- The forecast horizon is capped at 92 days. This is a flight planning tool, not an annual planning tool.
- On-target reach is not everyone who sees the ad. It reflects only people inside your specified age, gender and geography. Others see the ad but do not appear in the number.
- Seasonality. For auction-based formats, cost moves, so a forecast built in a quiet period will look optimistic for November and December — the same logic as in seasonal budget planning.
- Conversion forecasts are anchored to a specific budget level. Rates do not transfer cleanly to a budget twice the size.
- The model is built on historical, census and platform data and is periodically recalibrated. It is an estimate of a distribution, not a guarantee of an outcome.
Reach Planner answers “what order of magnitude of reach does this money buy.” It is poorly suited to “how much will I earn” — that answer comes from experiments and modelling, not a planner.
A working process, from plan to verification
Step 1. State the goal in reach terms, not click terms
Before opening the planner, decide how many unique people in the target audience should see the message and how many times. If the goal can only be phrased as “cheaper conversions,” you probably do not need reach buying — you need a performance campaign.
Step 2. Build the plan in Reach Planner
Set geography, demographics, period and budget. Compare two or three format mixes: skippable inventory usually buys more reach per dollar, while non-skippable buys higher guaranteed completion. Record the forecast reach and frequency — that becomes your reference point.
Step 3. Build the campaigns and group them
Create campaigns for your chosen formats, then add them to a video campaign group and set the objective. Check how many of them fall into the fully coordinated tier: if none do, the group objective is decorative.
Step 4. Reconcile plan against actuals in week two
Look at three things: unique reach, average frequency, and the frequency distribution (how many people got 1 exposure, how many got 5, how many got 15). An average of 3 with a long tail at 15+ is not fine — it means a narrow segment is being burned. Campaign-level capping and Target Frequency mechanics are covered in frequency capping and Target Frequency.
Step 5. Verify effect, not just delivery
Delivered reach is plan execution, not a result. Measuring the result needs Brand Lift, search lift and, at larger budgets, geo experiments and modelling. Click attribution structurally undercounts video — that is not a reason to call video useless, it is a reason to measure it with a different instrument.
Formats: what the money actually buys
Format choice matters more than the group objective — it determines both the cost of a contact and what counts as a contact at all.
| Format | What it buys | When it fits |
|---|---|---|
| Skippable reach inventory | Maximum unique people per dollar, low cost per contact | Broad reach goals, brand introduction |
| Non-skippable inventory | Guaranteed delivery of the full message, higher cost | A short, sharp message that cannot be truncated |
| Audio formats | Cheap incremental contact without visuals | Topping up frequency, supporting the main message |
| Connected TV inventory | Big screen, co-viewing | Brand work, broad household audiences |
Rule of thumb: do not build an entire plan on one format. Skippable buys breadth, non-skippable buys completion, audio tops up frequency cheaply. The combination usually produces a better reach-per-budget curve than any single format. But each format needs its own cut: a video designed to survive a skip at five seconds is built differently from a fifteen-second non-skippable.
A worked example: a six-week flight
- The brief. Reach 60% of the target audience in two regions at roughly three exposures per week over six weeks.
- The plan. In Reach Planner, set geography, demographics, period and budget; compare two format mixes. Record forecast unique reach, average frequency and cost per contact. Apply a seasonal adjustment if the flight lands in peak.
- The build. Create two or three reach campaigns for the chosen formats, add them to a group and set the frequency objective. Confirm the group contains coordinated campaign types, not just countable ones.
- Week one. Change nothing. The system is stabilising, and early conclusions are almost always wrong.
- Week two. Reconcile against plan: reach, average frequency, frequency distribution. If the average is below target and reach is behind too, the question is budget or an audience that is too narrow. If the average is on target but the tail runs past 12 exposures, the question is overlap between campaigns.
- Weeks three and four. One change at a time. Swapped creative? Leave the objective alone. Changed the objective? Leave the campaign set alone. Otherwise you will never know what worked.
- Weeks five and six. Run the effect measurement: a lift study, or a regional comparison if the budget supports a geo experiment.
- The report. Four numbers: planned reach, actual reach, planned frequency, actual frequency — plus the effect measurement on its own line. Do not blend plan execution with results; confusing the two is the source of most arguments about video budgets.
Common mistakes
| Mistake | Consequence | Do this instead |
|---|---|---|
| Group built only from Video Views and Demand Gen | The objective does not influence buying | Add reach campaigns the system can steer |
| Frequency of 6–7 on a long flight | Audience burnout, rising costs | Start at 2–3 and raise for a specific reason |
| Judging video on last-click CPA | Good campaigns get switched off as “inefficient” | Brand Lift, search lift, incrementality tests |
| Forecast built off-season, flight runs in peak | Actual cost exceeds plan | Apply a seasonal CPM adjustment up front |
| Same audience in every campaign in the group | Huge overlap, reach barely grows | Separate the audiences; check overlap in advance |
| Changing the group objective every other day | The system never stabilises | Give it 7–14 days of steady operation |
On audiences specifically: if segments across campaigns overlap by 70–80%, no amount of deduplication widens reach — you simply redistribute the same exposures more tidily. Separating segments by intent usually delivers more than tuning the group objective; how to build them is in custom segments and intent audiences.
Why your reach numbers will not match other systems
The familiar situation: Google Ads reports one reach figure, your agency’s report another, a panel study a third. That is not an error — they are different definitions.
- Unit of counting. The platform counts signed-in users and devices; panel studies count people; broadcast metrics count households. Co-viewing on connected TV is modelled, not directly observed.
- Audience boundary. On-target reach only counts people inside the specified age, gender and geography. Everyone outside those bounds saw the ad but is not in the number.
- Deduplication works inside the perimeter. Overlap between your YouTube campaigns is visible; overlap with an outside buy is not.
- Time period. Four-week reach is not the sum of four weekly reach figures — people repeat. Adding them is the single most common arithmetic error in reach reporting.
Practical conclusion: pick one source as your decision-making reference and compare it against itself. Reconciling three systems into one table takes days and almost never changes the decision.
Pre-launch checklist
- The goal is stated in reach and frequency terms, not clicks.
- A Reach Planner forecast exists and is recorded, with a seasonal adjustment applied.
- The group contains at least one fully coordinated reach campaign.
- Campaign audiences are separated and expected overlap is estimated.
- Creative is prepared per format rather than reused across all of them.
- The frequency target was chosen deliberately, not left at a default.
- An effect measurement method is assigned: lift study or geo experiment.
- Start date and every subsequent change are logged for later analysis.
- A hands-off period is agreed — at least one week.
- The final report format is defined: plan versus actual, plus a separate line for measured effect.
Who this is for
Groups plus Reach Planner earn their place if you run several concurrent YouTube campaigns against overlapping audiences; if the video budget has to be defended internally in recognisable media terms; if you have a dated flight with a reach commitment; or if you have already had complaints about repetition.
You can skip them if you run a single video campaign (campaign-level capping is enough); if video is purely a direct-response tool for you; or if budgets are small enough that deduplication disappears into the margin of error.
One sequencing note: get measurement right before optimising the buy. Planning reach while a chunk of your conversions and signals never reaches reporting produces a tidy plan and a disputed result. The base layer is clean tagging and reliable delivery, including first-party tag delivery through Google tag gateway. And if you also run automated campaigns, it helps to see how much budget lands on video inventory there — which is what Performance Max channel-level reporting now shows.
What comes next
Google has said coordinated frequency control across multiple YouTube line items is coming to Display & Video 360, without a date. The direction is clear enough: frequency management is moving from the campaign level up to the media plan level, closer to how broadcast buying has always worked. For advertisers that means “counting reach and frequency” stops being a specialist skill and becomes part of routine paid media work, alongside budget planning in Performance Planner. Format-level craft still matters too — see video ads and YouTube Shorts performance.
If the constraint is access and launch capacity rather than planning, that is a separate problem — see Google Ads agency accounts and the full PPC Rebels service list.
FAQ
How is a video campaign group different from a shared budget?
A shared budget pools money. A group shares a reach or frequency objective while budgets and creative stay separate.
Which campaigns does the system actually optimise inside a group?
Efficient Reach, Non-skippable Reach and Audio Reach are fully coordinated. Video Views, Demand Gen video, Target Frequency and reservation buys are counted but not steered.
What frequency should I start with?
Start at 2–3 weekly exposures on a long flight and raise it only for a specific reason — a short window, a narrow audience, or a message that needs repetition.
How accurate is Reach Planner?
It is a modelled estimate built on historical and census data and periodically recalibrated. Orders of magnitude are usually sound; precise values are not, especially in peak season.
Why is the forecast capped at 92 days?
The model anchors to the most recent comparable historical period. Beyond that window accuracy degrades enough that the forecast stops being useful.
What are co-viewing impressions?
An estimate of additional viewers watching the same connected TV screen. Without that correction, CTV reach is undercounted.
Can I group campaigns with different audiences?
Yes, and it is often better than running one audience everywhere: with heavy overlap, deduplication redistributes exposures rather than expanding reach.
How long before I judge the results?
At least 7–14 days of stable operation. Frequent objective changes prevent the system from settling and make periods incomparable.
How do I know frequency is too high?
Look at distribution rather than the average. If a meaningful share of the audience receives 10+ exposures a week, you are burning them even when the average looks modest.
Does a group replace campaign-level frequency caps?
No. A cap sets a ceiling inside one campaign; a group objective steers the pool toward an average. They solve different problems and can be used together.
How do I defend a video budget when last-click conversions look weak?
Show the effect with a different instrument: lift studies, geo experiments, modelling. Click attribution alone structurally undercounts upper-funnel work.
Do I need separate creative per format?
Ideally yes. Duration requirements and opening-seconds expectations differ between skippable and non-skippable inventory, and one cut rarely performs well in both.
