The account is open, tracking is installed, and the first question is usually how much to spend each month. Answering with a number right away tends to produce rework. Budget comes from market opportunity and unit economics, not from a target someone picked in a meeting. Know what one order or one lead leaves behind, and you know the CPA you can afford. Know how much demand is reachable, and you know whether more budget has anywhere to go.
Google Ads account setup and structure covers opening, verification, currency, MCC access, and naming. Those are the container. This article covers how much goes into the container and which target to aim at. It does not repeat the setup process. For the wider playbook, see the Google Ads optimization strategy guide.
Four Layers, Not One Number
The question of market size is usually reduced to one search volume figure. There are four layers, and each one loses volume.
| Layer | What it is | Main loss sources | How to estimate |
|---|---|---|---|
| Search volume | Estimated monthly queries from a tool | Non-commercial intent, wrong regions, duplicate queries, tool rounding | Keyword tool, use a range not a point |
| Addressable demand | The portion reachable within budget, bids, geography, language, match types, and policy | Budget caps, low bids, geo exclusions, policy limits | Small-budget tests, impression share, search terms |
| Clicks | Actual clicks delivered | Ad rank, creative appeal, competition, invalid traffic | Impressions times CTR, by campaign |
| Conversions | Orders or leads | Landing page, price, trust, sales follow-up, missing attribution | Clicks times conversion rate, reconciled with the backend |
Keyword tools estimate volume from historical search data to help with grouping and bids. The number is not a demand ceiling and definitely not a commitment. Treat it as a ceiling and the budget will be too large. Treat it as a promise and the forecast will miss. For finding demand-side terms and intent, see the keyword research and search intent guide.
Addressable demand takes another discount. Ten thousand monthly searches does not mean you can buy ten thousand clicks. Budget, bids, geography, language, match type, and category policy all cap the reachable share. Only a small live test shows that share. A tool cannot. For market splitting and currency handling, see the multi-country paid search guide.
The drop from clicks to conversions is larger still. A click is not a read, a read is not an inquiry, and an inquiry is not a deal. B2B adds a sales cycle, and DTC adds refunds and returns. A forecast writes each loss down instead of covering it with an industry average.
Calculate What You Can Afford
Unit economics come first. Work out what one order or one lead leaves behind, then derive the target CPA and ROAS.
| Input | Definition | Typical source |
|---|---|---|
| AOV | Average order value | Order system, finance |
| Gross margin | Revenue minus product cost, as a percentage | Finance |
| Refund and return rate | Share of revenue or orders refunded | Support, payments |
| Fulfillment cost | Shipping, packaging, payment fees | Logistics and finance |
| Repeat rate and LTV | Future contribution from the same customer | CRM, cohorts |
| Sales cycle | Time from lead to deal | CRM |
| CVR | Click-to-conversion or click-to-lead rate | Analytics, CRM |
| CPC | Cost per click | Ads platform, live tests |
Start with contribution per order.
Gross profit per order = AOV times gross margin rate
Net contribution = gross profit minus fulfillment cost, or gross profit times (1 minus refund rate)
Then derive the break-even and target values.
Break-even CPA = net contribution per order
Target CPA = break-even CPA minus desired profit per order
Break-even ROAS = AOV divided by break-even CPA
Target ROAS = AOV divided by target CPA
An illustrative example, with numbers chosen to show the formula rather than to set a benchmark. AOV 120, gross margin 45 percent, refund rate 8 percent, desired profit 10 per order. Gross profit is 54, net contribution about 49.7, target CPA about 39.7, break-even ROAS about 2.4, target ROAS about 3.0. If CPC is 1.2 and CVR is 3 percent, cost per conversion is about 40, right on the target line. There is almost no room for error.
Lead-generation businesses use a different formula.
Break-even cost per lead = contract value times gross margin times lead-to-deal rate times (1 minus churn rate)
With a contract value of 20,000, gross margin of 55 percent, and a 12 percent lead-to-deal rate, break-even cost per lead is about 1,320. At a 20 percent target margin, the target cost per lead is about 1,056. A sales cycle of three to six months also has to survive the cash flow, not just the final profit.
LTV can raise what you can afford, but keep two ceilings. The cash ceiling is first-order or first-year payback. The LTV ceiling is long-term contribution. Using LTV to justify a loss on the first order only works when repeat purchase data actually exists.
Three Input Tables
B2B leads, DTC orders, and high-ticket businesses need different inputs. One table cannot serve all three.
B2B Leads
| Input | What to fill in | Use |
|---|---|---|
| Contract value | Annual or per-deal value | Value per lead |
| Gross margin | After delivery cost | Affordable cost per lead |
| MQL to SQL | Marketing lead to sales-accepted rate | Lead quality |
| SQL to win | Sales-accepted to signed rate | Full-funnel conversion |
| Sales cycle | Average days to close | Cash flow and observation window |
| Target cost per lead | Derived from value and conversion | Media target |
| Payback period | Time to recover acquisition cost | Budget ceiling |
DTC Orders
| Input | What to fill in | Use |
|---|---|---|
| AOV | Average order value | Revenue per order |
| Product cost | Including packaging | Gross profit |
| Fulfillment and payment | Shipping, fees | Net contribution |
| Refund rate | Share of orders or revenue | Revenue correction |
| Repeat rate | 90-day or 12-month repurchase | LTV judgment |
| CVR | Click to order | Clicks needed |
| CPC | Live or ranged | Budget |
High-Ticket and Long Cycles
| Input | What to fill in | Use |
|---|---|---|
| Deal size | Contract or project value | Value per lead |
| Offline close rate | Lead to signed deal | Affordable cost |
| Sales cycle | In months | Longer observation window |
| Cost per qualified lead | Qualified, not raw form fills | Avoid low-quality lead distortion |
| Cash tolerance | First-year acquisition spend | Budget ceiling |
| Other channel contribution | Organic, events, referrals | Avoid attributing everything to ads |
All three tables share one starting point. Work out what a lead or an order leaves behind before discussing budget. They differ in emphasis. B2B looks at sales cycle and lead quality, DTC looks at refunds and repeat purchase, and high-ticket looks at offline close rates and cash tolerance.
Derive Budget from the Target
With a target CPA and an estimated CVR, work back to clicks and budget.
Clicks needed = target conversions divided by CVR
Budget = clicks needed times CPC
Test budget = decision-threshold conversions times target CPA
The threshold is a team decision rule, not a substitute for statistical power. Pick 25 conversions or four weeks, whichever comes first. At 25 conversions and a target CPA of 40, the test budget is about 1,000. With CPC at 1.2 and CVR at 3 percent, clicks needed are about 833 and spend is about 1,000. The three numbers reconcile.
When CVR is unknown, do not use a point estimate. Use a range, such as 1.5 to 4.5 percent, and calculate budget and CPA at both ends. Check whether the worst case is survivable. The range should come from your own analytics, comparable pages, or a small test, not from intuition.
Budget is also a time decision. At 100 per day, a 1,000 test takes at least ten days. At 30 per day, it takes over a month. Too short and the data is thin. Too long and the adjustment arrives late. Set budget, daily spend, and observation window together.
Forecast as a Range
A forecast table should not carry a single number. Use at least three columns, low, base, and high, with the source of every input.
| Input | Low | Base | High | Source |
|---|---|---|---|---|
| CPC | 0.9 | 1.2 | 1.6 | Small test or platform range |
| CVR | 1.5% | 3.0% | 4.5% | Comparable pages and past campaigns |
| Target CPA | 35 | 40 | 45 | Derived from unit economics |
| Monthly budget | 800 | 1,200 | 1,800 | Target conversions times CPA |
Write outputs as ranges too, monthly conversions, CPA, and ROAS. The forecast manages expectations and triggers reviews. It is not a promise to a manager.
Calibration Cadence
Forecasts will be wrong. What matters is knowing when to recalibrate and what to recalibrate.
In week two, check data quality only. Confirm tracking is complete, search terms are relevant, and CPC and CTR are in a plausible range. CPA is still unstable, so do not draw conclusions.
In week four, run the first formal calibration. Put actual CPC, CVR, and conversion counts into the formulas, recalculate CPA and ROAS, and compare against the forecast range. Identify which layer caused the gap, whether clicks were short, conversion rate was low, or AOV and refunds differed from the assumptions.
In week eight, add backend data. DTC checks refunds and repeat purchase. B2B checks lead quality and sales progress. High-ticket checks offline close rates. Only then can you judge whether the unit economics hold, rather than whether the platform CPA looks good.
Record every assumption change, who made it, and why. The next market forecast starts from that record.
Common Misconceptions
Search volume is the demand ceiling. Tools estimate. Addressable demand is capped by budget, bids, geography, and policy.
The platform forecast is a commitment. It is built on historical data and assumptions, and it needs calibration.
Set ROAS without looking at margin. Revenue is not profit. A low-margin business can lose money at the same ROAS.
Ignore refunds and sales cycles. DTC refunds and long B2B cycles change the real acquisition cost.
Make the test budget too small. Too few conversions means the next decision is a guess again.
Optimize CPA before tracking is reliable. If conversion definitions and attribution are unstable, the CPA is not comparable. That work belongs to the conversion tracking and attribution guide.
Related Reading
- Google Ads account setup and structure, opening, verification, currency, and naming.
- Bidding strategy and budget pacing, executing and adjusting target CPA and ROAS.
- Conversion tracking and attribution, the source of the forecast inputs.
- Google Ads cost and budget, cost structure and budget context.