Google Ads Media Plan 2026: Costing the Channel Before the First Click
“How much budget do we need?” gets asked before the account has produced a single click, and no honest answer is one number. A Google Ads media plan built before launch has five layers, and Google supplies only two of them. The other three are yours — and that is where the entire forecast error lives. Below: what Keyword Planner actually forecasts (spoiler: not conversions), why Performance Planner is unusable pre-launch by its own eligibility rules, the arithmetic that connects demand to money, how to compute the minimum budget at which a campaign can even learn, and how to hand a client three scenarios instead of one attractive number.
Two forecasting tools, and only one works before launch
Google Ads has two forecasting tools and they get confused constantly. The difference is structural.
| Keyword Planner | Performance Planner | |
|---|---|---|
| Needs a running account | No | Yes |
| What it forecasts | Clicks, impressions, cost, CTR, average CPC | Campaign performance under budget and target changes |
| Conversions in the forecast | No | Yes, against the chosen goal |
| Horizon | Daily (derived from weekly data) | Monthly and quarterly |
| When to use it | Pre-launch, when semantics change | After a month of spend, and onwards |
Performance Planner’s eligibility rules for Search, Shopping (Standard) and Performance Max shut the door on a new account outright: no bid strategy change in the last 10 days, at least 3 recent clicks, at least 10 impressions, at least 1 conversion or conversion value, and spend greater than zero in the last 17 days. App campaigns are stricter still — running for 10 days or more, at least 10 conversions in the last 10 days, tCPI or tCPA bidding, and not limited by budget. Demand Gen has its own grid: no bid strategy change in the past 7 days, at least 14 impressions in 7 days for tCPA, Maximize Conversions and Maximize Clicks, at least 5 conversions in 7 days for tROAS and Maximize Conversion Value, and no experiments, product feeds, flighting or mixed bidding strategies.
What remains before the first click
So for a plan written before the first click, exactly one source remains: the Keyword Planner forecast. We cover how to assemble the keyword set it runs on separately in building a keyword set you can trust; here we work from a list that already exists.
One more detail worth filing away: as of 9 March 2026, Performance Planner no longer supports planning for Display and Video campaigns, or plans whose primary metric is impression share, top impression share or absolute top impression share. As a result, you can no longer view or edit those plans. What remains supported: Search, Shopping (Standard), Performance Max, Demand Gen and App.
What the Keyword Planner forecast actually returns
Five metrics, all expressed per day:
- Clicks — how many clicks the ad might receive in a day if the keyword triggers it.
- Cost — the average amount you might spend per day on that keyword.
- Impressions — how often the ad might show in a day.
- CTR — forecast clicks divided by forecast impressions.
- Average CPC — the average amount you might pay per click.
However, conversions, conversion rate and cost per conversion are not on that list. This is the single most important fact about the forecast: Google estimates traffic, not outcomes. Any line in a media plan reading “we expect 120 leads” is your assumption, not the platform’s forecast, and you should label it that way.
The underlying data is thinner than it looks, too. In Google’s own phrasing, forecasts are based on one week of data, averaged to produce daily figures. Forecast impressions account for your bid, budget, seasonality and historical ad quality — while the search volume statistics account for none of that. In other words, these are two different numbers in one interface, and you must not mix them.
Meanwhile, average monthly searches is an average over a 12-month period and shown only for exact match. Traffic forecasts, unlike it, do take match types into account. Pulling volume from the statistics column to plan a broad-match campaign is a methodological error, not a rounding one.
The five layers of a Google Ads media plan
A structure that makes visible where the data ends and your hypotheses begin.
| Layer | Source | Whose risk |
|---|---|---|
| 1. Demand: how many queries exist | Keyword Planner, 12-month average | |
| 2. Reachable share at a given bid | Forecast impressions at that bid | |
| 3. Clicks and cost | Forecast clicks and CPC | |
| 4. Site conversion rate | Your assumption | Yours |
| 5. Economics: allowable CPA | Your unit economics | Yours |
The arithmetic connecting the layers fits in four lines:
- Clicks = Impressions × CTR
- Cost = Clicks × CPC
- Conversions = Clicks × CR (CR is yours)
- CPA = Cost ÷ Conversions
Plus the line that usually decides the project: allowable CPA = gross margin per order × target marketing share of margin. If the allowable CPA sits below the forecast CPA at a realistic CR, the channel does not close — and therefore it needs to be visible in a spreadsheet before launch, not two months in.
Worked example with illustrative numbers
Take a B2B service with an average deal of $800 at 40% margin — $320 of gross profit per deal. Target marketing share of margin is 30%, so the allowable cost per deal is $96.
At a $1.60 bid the forecast returns 1,800 impressions a day at 5.5% CTR, so roughly 99 clicks a day at an average CPC of $1.27 — about $126 a day, or $3,780 a month.
Now the layer Google does not compute. Assume a 3% site conversion to enquiry and a 25% enquiry-to-deal rate.
| Scenario | Site CR | Leads/mo | CPL | Deals | Cost per deal | Verdict |
|---|---|---|---|---|---|---|
| Pessimistic | 1.5% | 45 | $84 | 11 | $344 | Does not close |
| Base | 3.0% | 89 | $42 | 22 | $172 | Does not close |
| Optimistic | 4.5% | 134 | $28 | 33 | $115 | Borderline |
The conclusion matters more than the numbers: even the optimistic scenario misses the $96 allowable. So the pre-launch conversation is not “give us a bigger budget” but three specific levers — raise the enquiry-to-deal rate, cut CPC with tighter semantics, or revisit the allowable marketing share. After all, that is what a media plan is for: it changes the agenda before the money moves.
Every figure above is illustrative. Substitute your own.
Minimum budget: what it takes for a campaign to learn
A calculation almost always missing from media plans. Automated bidding needs conversions, and a budget that yields five a month will never exit learning — you pay for traffic and get no optimisation.
The formula: minimum budget = (target conversions ÷ CR) × CPC.
For the example above, targeting 30 leads a month at 3% CR: 30 ÷ 0.03 = 1,000 clicks × $1.27 = $1,270 a month. That is the floor at which a discussion about automated bidding is meaningful at all. If the client offers $500, the honest answer is not “let’s try” but “at that level the campaign stays manual and collects data; do not expect optimisation”.
Keep this as its own line in the plan. It turns an abstract budget argument into a concrete condition.
Nine places the forecast is wrong — by Google’s own admission
Google lists the limitations of its own forecast, and that list is more useful than any accuracy claim.
- These are approximations. The documentation uses that word and offers no accuracy percentage and no confidence interval.
- For new advertisers, forecasts are based on historical average data for all advertisers. That describes the market, not you.
- Search Network only. Also, the forecast excludes the Display Network.
- Small geographies are less accurate. Stated plainly.
- Overlapping keywords degrade accuracy. Noted separately: Keyword Planner does not look at keywords across campaigns, so the tool does not model cannibalisation at all.
- Match types only affect traffic forecasts. Likewise, historical search volume covers exact match only.
- Search partners are reflected with adjustments for differing site policies.
- One week of data underpins it. Any spike or dip in that week carries into the daily average.
- Bid ranges cover the last 30 days, with the low and high ends at roughly the 20th and 80th percentiles. That is a picture of the market, not your future CPC.
Point five deserves a comment. If you are planning Search, Shopping and Performance Max together, the search forecast will overstate: the other campaigns will take part of that traffic. The adjustment is manual, and a conservative 15–25% is a reasonable starting haircut — an estimate, since Google publishes no coefficient.
Which bid to feed the forecast
By default Keyword Planner inserts a bid from your previous plan, or the average manual max CPC across manual-CPC ad groups in the account, or the average across manual-bidding ad groups in the same currency. For a brand-new account none of those sources means anything.
A workable sequence:
- Read the top-of-page bid range (low / high) — the 20th and 80th percentiles of the market over 30 days.
- Treat the low end as the cautious scenario and the high end as the aggressive one.
- Pull the forecast three times: low, midpoint, high.
- Put all three side by side. The spread between them is the plan’s real uncertainty.
You can change the bid either by selecting a point on the graph or by typing into the field at the top of the page. After launch this same question moves from forecasting to bid, budget and target simulators, which run on your account’s actual data.
Six common mistakes in a media plan
Taking volume from the “Average monthly searches” column and multiplying by CTR. That column is a 12-month average for exact match, blind to bid, budget and quality. Clicks come from the forecast column, not the statistics one.
Dropping in an industry-average conversion rate. Benchmarks are reference points, not data. With no first-party number, three scenarios beat one borrowed figure.
Forgetting tax and fees. Note that the budget in the plan and the amount leaving the card are different numbers.
Ignoring conversion lag. In B2B an enquiry becomes a deal weeks later. A monthly plan that puts spend and deals in the same column systematically understates CPA in month one and overstates it in month two.
Planning broad match on exact-match statistics. How match types changed by 2026, and what is left of exact, is the topic of keyword match types and close variants.
Budgeting without negatives. The forecast counts all traffic on a keyword, including the traffic you will exclude in week one. A 10–20% haircut at launch is sensible caution; details in negative keywords and search term analysis.
How to present the Google Ads media plan
A single number in a media plan is a promise you do not control. Three scenarios are a frame for a decision. The minimum set worth showing:
| Row | Contents | Source |
|---|---|---|
| Demand | Monthly queries across the core set | Keyword Planner |
| Traffic | Impressions, clicks, CTR, CPC at three bids | Forecast |
| Spend | Three budget scenarios | Forecast |
| Conversion | Three CR values, labelled as assumptions | You |
| Outcome | Leads, CPL, deals, cost per deal | Calculation |
| Threshold | Allowable CPA from unit economics | Client |
| Learning floor | Budget below which automated bidding will not work | Calculation |
| Caveats | Which lines are Google’s forecast and which are hypotheses | You |
That last row protects both sides. When reality diverges from the plan a month later, the conversation is about which hypothesis failed, not about why you promised something.
A checklist before you send it
Ten items that separate a media plan from a spreadsheet of attractive numbers. Working through them takes fifteen minutes and removes most future disagreements.
- Every row is labelled as either a Google forecast or your own assumption.
- Traffic volume came from the forecast column, not from average monthly searches.
- You pulled the forecast three times — at the low, mid and high bid from the range.
- You applied a haircut for the negatives you will add in week one.
- You applied a haircut for overlap with other campaigns, if any are planned in parallel.
- Conversion rate is shown as three values, not one.
- You calculated the minimum budget for automated bidding to exit learning.
- The allowable CPA comes from the client’s unit economics, not from guesswork.
- You accounted for conversion lag: in month one, spend and deals do not line up in time.
- There is a line for review and certification timelines if the category requires them.
An unchecked item is not a reason to withhold the plan — it is a reason to write one line next to it naming what is missing. An acknowledged assumption costs far less than an undiscovered error.
The first four weeks: reconciling plan against reality
| Week | What to check | Normal variance |
|---|---|---|
| 1 | CPC and impressions | Up to 40%: bids and volume swing at launch |
| 2 | CTR | Up to 30%: driven by your ads, not the forecast |
| 3 | Site CR | This is your hypothesis — it is the thing under test |
| 4 | CPL and monthly spend | Up to 25% is workable; more means rebuilding the plan |
The ranges above are practical reference points for month one, not Google standards. The purpose of reconciliation is not to hit the forecast but to identify which layer was wrong: demand, auction, ads or site.
After a month the second tool joins in. By then, the account usually clears Performance Planner’s eligibility bar, and planning moves onto your own data — covered in building a Q4 budget you can defend. Seasonal demand in your own category is easiest to read in the Insights page demand forecasts.
What neither forecast will ever show you
- Your site’s conversion rate. Not Keyword Planner, not the market — only your own test.
- Competitor behaviour. Performance Planner simulates auctions from the last 7–10 days, not rivals’ plans for next quarter.
- Cannibalisation between campaigns. Keyword Planner explicitly does not look across campaigns.
- Lead quality. A forecast counts events, not money.
- Review timelines. If the category needs certification, launch can slip by weeks — that belongs in the plan as a line, not as a surprise. What to do about it is in Google Ads Policy Manager and handling disapprovals.
The media plan is the third of three control layers a human keeps in an automated account: what you are allowed to show, what the user is guaranteed to see — that is pinning RSA headlines — and what the whole thing costs. Automation optimises inside the frame; you draw the frame.
When the plan is blocked not by arithmetic but by platform access — a category that requires a verified advertiser, for instance — it is worth looking at agency Google Ads accounts and the full list of PPC Rebels services. The long-form version of budget and account-structure methodology lives in the Google Ads course.
FAQ
Does Keyword Planner forecast conversions?
No. It returns clicks, impressions, cost, CTR and average CPC, all on a per-day basis. Conversions, conversion rate and cost per conversion are assumptions you add yourself.
Can I use Performance Planner before launch?
No. Search, Shopping (Standard) and Performance Max each require at least 3 recent clicks, 10 impressions, 1 conversion and spend above zero in the last 17 days, with no bid strategy change in the last 10 days.
What data is the Keyword Planner forecast built on?
One week of data, averaged into daily figures. Forecast impressions take your bid, budget, seasonality and historical ad quality into account.
Why can’t I use “Average monthly searches” to calculate clicks?
It is a 12-month average shown only for exact match, with no regard for bid, budget or quality. Traffic forecasts are a separate column, and they do account for match types.
How accurate is the forecast for a brand-new account?
For new advertisers the forecast is based on historical average data across all advertisers, so it describes the market rather than you. Google calls the figures approximations and publishes no accuracy percentage.
How do I calculate the minimum budget for automated bidding?
(Target conversions ÷ conversion rate) × CPC. At 30 conversions, 3% CR and $1.27 CPC, that is 1,000 clicks and about $1,270 a month.
Does the forecast account for cannibalisation between campaigns?
No. Google states that Keyword Planner does not look at keywords across campaigns, so overlap between Search, Shopping and Performance Max has to come off manually.
Which bid should I enter into the forecast?
Run it three times — at the low end of the bid range, the midpoint and the high end. The range is built from the last 30 days and corresponds roughly to the 20th and 80th percentiles.
What changed in Performance Planner on 9 March 2026?
Google removed planning for Display and Video campaigns, as well as plans using impression share, top impression share or absolute top impression share as the primary metric. As a result, you can no longer view or edit such plans.
Which campaign types does Performance Planner support now?
Search, Shopping (Standard), Performance Max, Demand Gen and App. Display and Video are no longer listed in the documentation.
How often do Performance Planner forecasts refresh?
Daily, based on the last 7–10 days and adjusted for seasonality. The planner simulates auctions over that period, factoring in seasonality, competitor activity and the landing page.
What variance between plan and actuals is normal?
In month one, up to 25% on spend and CPL is workable; CPC and impressions often diverge more at launch. These are practical reference points, not Google standards.
