Break-even CPA & target ROAS math for service businesses — set Google Ads targets that finance will sign
How service SMBs calculate break-even CPA and allowable media CPA from job value, margin, and close rate — plus when to use Target CPA vs break-even ROAS math — so Smart Bidding targets stop being round numbers from a YouTube video.
Direct answer — what is break-even CPA for Google Ads?
Break-even CPA is the highest cost per conversion you can pay and still make zero profit on that conversion path — given your job value, contribution margin, and lead-to-close rate. For service businesses, build it from economics, not from last week’s Google Ads average. Then set Target CPA near recent efficient CPA, at or below a finance-approved allowable CPA (break-even minus the profit cushion you need). If your conversion is a soft form fill, the math is fiction until the conversion definition matches sales reality.
This insight owns service-business CPA / ROAS economics for bid targets. Strategy selection: Maximize Conversions vs Target CPA. Value-based ecommerce floors: Break-Even ROAS calculator and low-volume Target ROAS. Full bidding tree: Smart Bidding SMB playbook.
If finance will not sign the inputs, do not put the output into Target CPA.
Why service SMBs get this wrong
Round-number targets. “Let’s do £50 CPA” because it sounds disciplined. No link to job value or close rate.
Wrong conversion in the denominator. Math uses qualified booked jobs; Google optimizes form fills. Allowable CPA on forms can look “generous” while every form still loses money after sales time (bad conversion data).
Blended brand + non-brand. Brand CPA is cheap; non-brand is expensive. One Target CPA averages a lie that starves growth or overpays junk.
Copying ecommerce ROAS. Services without revenue values in Ads should usually use count-based Target CPA on a qualified event — not a costume Target ROAS on $1 lead values (low-volume tROAS).
Ignoring capacity. Allowable CPA assumes you can fulfill the jobs. If the diary is full, the constraint is ops — not a tighter tCPA.
Agency theatre. Targets change every Monday without a signed one-pager. Learning never ends; nobody can say whether economics or thrash caused the result (learning phase).
Confusing CPL with CAC. Cost per lead is not customer acquisition cost. If you celebrate CPL while ignoring close rate and contribution, you will scale unprofitable lead volume.
The service break-even CPA formula
Use this operating model:
Contribution per closed job = Average job revenue × contribution margin %
(Use contribution after direct job costs — labour, materials, subcontractors you treat as variable. Align VAT/tax treatment with finance.)
Break-even CPA on a closed-job conversion ≈ contribution per closed job
(At that CPA you pay all contribution away to media — zero leftover for overhead/profit unless overhead is already excluded by design.)
Most Google Ads “conversions” are leads, not closed jobs. So:
Break-even CPA on a lead conversion ≈ contribution per closed job × lead→close rate
(Optional: subtract sales cost per lead if SDRs/closers are material and you want media to clear that too.)
Allowable CPA = break-even CPA × (1 − required profit cushion)
or break-even CPA minus a fixed £/$ profit per lead you insist on. Pick one method and document it.
Algebra you can paste into a sheet
contribution_per_job = avg_revenue * margin_pct
be_cpa_closed = contribution_per_job
be_cpa_lead = contribution_per_job * close_rate
allowable_cpa = be_cpa_lead * (1 - cushion_pct)
Example cushion: 0.20 for 20%. If close_rate is 0.25 and contribution is 400, break-even lead CPA is 100; allowable at 20% cushion is 80.
Worked example (UK trades)
- Average job revenue: £800
- Contribution margin: 50% → £400 contribution per closed job
- Lead→close (qualified lead to paid job): 25%
- Break-even CPA on qualified lead: £400 × 0.25 = £100
- Finance wants ~20% cushion on that path → allowable ≈ £80 Target CPA ceiling for that conversion definition
If Google’s primary is “any form submit” with a 10% close rate instead of 25%, break-even on that softer event is £400 × 0.10 = £40 — not £100. Using £80 Target CPA on soft forms overpays junk.
Worked example (US home services)
- Average ticket: $1,200
- Contribution margin: 45% → $540 per closed job
- Close rate from sales-accepted lead: 30%
- Break-even CPA on accepted lead: $540 × 0.30 = $162
- Allowable with 15% cushion ≈ $138
Call-heavy shops: if the conversion is a 90-second call, use the close rate from that call type, not from web forms.
Worked example (multi-service line)
Emergency callout jobs: $900 contribution, 40% close from qualified → break-even lead CPA $360.
Planned maintenance: $200 contribution, 20% close → break-even lead CPA $40.
One blended Target CPA of $150 will overpay maintenance leads and under-buy emergency. Split campaigns or use value-based bidding with honest stage values — do not average incompatible economics.
Map the math to the conversion action
| Conversion you optimize | Close-rate input must be | |-------------------------|---------------------------| | Form thank-you | Form → closed job | | Call ≥ duration | Call → closed job | | Offline Qualified (OCI) | Qualified → closed job | | Offline Won / booked job | ≈ 100% (already closed) — CPA break-even ≈ contribution per job |
Offline won as primary: Target CPA ≈ allowable cost per booked job (contribution minus cushion). That is the cleanest service economics — when volume supports Smart Bidding (OCI playbook).
Primary vs secondary must match the math inputs (primary vs secondary rules). Changing the primary without rewriting allowable CPA is how “tCPA stopped working” myths start.
Quick integrity test
Ask: “If Google delivered one more of this conversion for exactly our Target CPA, would finance be happy after close rate and contribution?” If the room hesitates, the conversion definition or the number is wrong.
Allowable CPA vs Target CPA vs Maximize Conversions
Allowable CPA is a finance ceiling.
Target CPA is the Google Ads bid strategy input.
They are related — not identical.
| Situation | Bid approach | |-----------|----------------| | Clean primaries, thin history, unknown efficient CPA | Maximize Conversions first; compute allowable in parallel | | 30+ qualifying conv/mo, history near allowable | Target CPA near recent efficient CPA, at or under allowable | | History CPA already above allowable | Do not “fix” with fantasy tCPA — fix CR, offer, geo, or accept lower volume | | Soft primaries | Rebuild measurement before any target math |
Full switch rules: Max Conv vs tCPA. After target changes, respect the learning phase.
Setting tip: if trailing 30-day CPA on the correct conversion is £95 and allowable is £80, do not jump to £80 overnight. Step toward it (e.g. £92 → £88 → £84) while watching volume and MER — or fix conversion rate so £80 becomes reachable.
Portfolio tip: do not force five campaigns with different close rates into one portfolio Target CPA unless finance accepts blended economics explicitly.
When services should use break-even ROAS instead
Use ROAS math when Ads receives honest conversion values (purchase revenue, offline won revenue, or tiered lead values that finance trusts).
Rough ecommerce-style relation: break-even ROAS ≈ revenue ÷ contribution per order (tooling handles shipping, fees, returns). Run the Break-Even ROAS calculator for that path.
For services with equal fake lead values, “Target ROAS” is Target CPA in costume — prefer Target CPA on qualified counts until offline values exist.
Hybrid: phone-won jobs with revenue import → value-based bidding possible; web quote forms stay count-based until valued.
Translation check: if you assign £80 expected contribution as conversion value on a qualified lead, Maximize Conversion Value / Target ROAS should be judged against break-even ROAS math — not against a vanity 800% target.
Inputs finance must approve (one-pager)
Before any Target CPA goes live, write:
- Average job revenue (and by service line if margins differ)
- Contribution margin definition
- Conversion definition in Google Ads (exact action name)
- Lead→close rate source (CRM, last 90 days)
- Break-even CPA and allowable CPA
- Brand vs non-brand (separate targets if CPAs differ materially)
- VAT/tax and currency notes
- Review date (quarterly or when offer/pricing changes)
If sales and marketing disagree on (3) or (4), stop. The number in Google is politics until definitions match.
Print the one-pager into the Ads change log the day you set Target CPA. Future you will need it.
Multipliers and traps that inflate “allowable”
LTV hope. Using three-year LTV to justify 3× higher CPA without measured retention is how accounts scale into losses. Only include repeat purchase when you have data — same discipline as the ROAS tool’s LTV toggle.
Overhead double-count. If contribution already excludes overhead, do not subtract full company overhead again into CPA unless finance designs a fully loaded CAC model deliberately.
Seasonality. Winter HVAC close rates ≠ summer. Recalculate allowable by season if mix shifts.
Sales capacity. Close rate collapses when the phone is unanswered — that is an ops KPI, not a bid strategy bug. Tightening tCPA will not hire technicians.
Aggregator leads. Close rates from Angi-style leads often differ from Google Search leads — do not reuse one close rate across sources.
Refunds and no-shows. If 10% of “closed” jobs cancel, your effective contribution and close rate need that haircut — or you will set allowable CPA on fantasy completes.
Brand, geo, and campaign structure implications
Separate brand Search from non-brand before applying one Target CPA. Brand can clear allowable easily while non-brand sits above break-even — blending hides the truth.
Geo: high-CPC metros may need higher tickets or better CR to clear the same allowable; otherwise shrink geo or raise prices — bidding cannot invent margin.
PMax: same economics, worse transparency. Do not set PMax tCPA/tROAS from vanity; sequence with the PMax playbook after measurement is honest.
Device and call mix: if mobile drives calls that close better than desktop forms, your blended CPA target may need path-aware measurement (multi-path tracking) before you trust one number.
US and UK notes
US: Ticket sizes and close rates vary hard by trade and metro. Multi-location franchises need per-market or per-offer allowables — HQ “national CPA” often starves expensive cities. Call conversions need duration-qualified close rates.
UK: Quote VAT treatment with finance (revenue and margin inputs). LSA vs Search: do not casually apply Search allowable CPA to a different lead product. Consent does not change the formula — it can change volume and match rates.
Both: Report MER and CAC payback beside Google CPA (metrics that matter). Platform CPA is a means; cash is the end.
Implementation sequence (one week)
Day 1: Pull 90 days CRM — revenue, margin proxy, close rates by source.
Day 2: Inventory Google primary conversion actions; align definitions with sales.
Day 3: Compute break-even and allowable CPA per money conversion; finance signs the one-pager.
Day 4: Compare trailing Ads CPA on that conversion to allowable; decide Max Conv vs step-down tCPA.
Day 5: Screenshot baselines; change targets carefully; calendar a 14-day review (learning phase).
Ongoing: Recalc when pricing, close rate, or conversion definition changes.
Run Google Ads Health Score and Tracking Trust Score if you cannot trust the conversion column yet.
What “good” looks like after 30 days
Trailing CPA on the signed conversion sits at or under allowable most weeks; volume is acceptable to ops; MER is stable or up; nobody is arguing about a Target CPA that was never tied to contribution. If CPA is under allowable but MER is down, look at creative waste, brand cannibalization, or non-Google spend — not only Search tCPA.
Spreadsheet starter (copy these columns)
Build one tab finance can audit:
| Column | Example | |--------|---------| | Service line | Emergency / Planned | | Avg revenue | 800 | | Margin % | 50% | | Contribution / job | 400 | | Google conversion name | Offline – Qualified | | Close rate from that stage | 25% | | Break-even CPA | 100 | | Cushion % | 20% | | Allowable CPA | 80 | | Trailing 30d Ads CPA | 95 | | Gap | 15 over allowable | | Action | Step tCPA down / fix CR |
Update the trailing CPA weekly; update economics when price or close rate moves. If the sheet does not exist, Target CPA does not deserve to exist. Owners who refuse the sheet are asking media to invent profitability. Put the allowable CPA in the Ads change log the day you set it.
FAQ
What is break-even CPA in Google Ads?
It is the maximum cost per conversion at which media cost consumes all contribution from that conversion path — zero profit left. For leads, multiply contribution per closed job by lead→close rate.
How do I calculate Target CPA for a service business?
Compute break-even CPA from job value, margin, and close rate on the same conversion Google optimizes; set allowable CPA with a profit cushion; set Target CPA near recent efficient CPA and at or below allowable — not 40% below history on day one.
Is break-even CPA the same as Target CPA?
No. Break-even/allowable is finance math. Target CPA is the bid strategy setting. Target should respect allowable, but volume and learning may require starting closer to historical CPA.
Should service businesses use break-even ROAS instead?
When conversion values in Ads are honest revenue (or trusted proxies). Otherwise use break-even CPA / Target CPA on qualified counts. Use the Break-Even ROAS calculator for value-based paths.
What close rate should I use?
The close rate from the same stage as your Google conversion action, from recent CRM data — not a national benchmark from a blog.
Why did CPA look fine but we still lost money?
Often soft conversions, ignored sales cost, wrong close rate, or brand/non-brand blending. Reconcile to contribution and MER.
Can I use LTV to raise allowable CPA?
Only with measured repeat purchase data. Hoped-for LTV is how targets become fiction.
How often should I refresh break-even CPA?
At least quarterly, and whenever pricing, offer, close rate, or conversion definitions change.
What if my historical CPA is already above allowable?
Fix conversion rate, offer, geo, or creative — or accept less volume. A fantasy Target CPA below history usually kills spend without creating margin.
Does this replace Smart Bidding strategy choice?
No. Economics set the ceiling and sanity check. Strategy choice (Max Conv vs tCPA vs value bidding) is covered in the Smart Bidding playbook and sibling insights.
What if close rate is unknown?
Do not invent Target CPA. Pull 90 days of CRM outcomes first, or stay on Maximize Conversions while you instrument stages. A guessed close rate is how allowable CPA becomes fiction.
Should brand and non-brand share one Target CPA?
Only if finance accepts blended economics. Usually separate — brand will otherwise subsidize a target that non-brand cannot clear.
Related frameworks
- Break-Even ROAS calculator — margin math for value-based targets
- Smart Bidding SMB playbook — strategy tree
- Smart Bidding readiness scorecard — 36 checks before targets
- Maximize Conversions vs Target CPA
- Target ROAS when volume is low
- SMB marketing metrics that matter
- ROI Calculator — benchmark context beside unit economics
- Free Google Ads audit — we will tell you whether your Target CPA is economics or a round number
