Ad costs: how to calculate ROAS and the maximum cost per order and customer
The ads dashboard said every dollar brings back 4. The bank said something else
In the pricing guide we met Majed, the perfume store owner who thought he made $36 on every bottle, then found out he keeps only 25 cents
A week later Majed opened Snapchat and saw a return on ad spend of 3.9. TikTok showed 4.0. So he told himself: every dollar I spend brings back 4, why not double the budget? He doubled it, and by the end of the month his loss had almost doubled too
This guide explains how to calculate ad costs the right way, from the simple math everyone knows to the math growth managers use in large companies. We apply it to a store, a service business and a subscription business, with calculators you can try on your own numbers
Majed, Sara and Khaled are illustrative characters with costs close to market prices. The eBay, Uber, Airbnb and Procter & Gamble numbers are real and published, with links in the text. Technical terms link to pages from Google, Meta and others that explain them
An ad is a price you pay for an order
Most owners see ads either as an expense to cut or an investment to grow. A more accurate way is to see an ad as the price you pay to buy an order or a customer
And if it is a price, only one question matters: is the price you pay for an order lower than what the order leaves you? If it is lower, every ad dollar makes money and a bigger budget makes sense. If it is higher, every ad dollar loses money, and a bigger budget grows the loss no matter how good the dashboard looks
The maximum cost per order is what an order leaves you after all its costs except the ad. It is the most you can pay for ads on one order before you start losing money
The whole guide is built on this idea, and we apply it three ways because the "order" changes by sector: a store buys an order, a service business buys an inquiry that may turn into a contract, and a subscription business buys a customer who pays in installments
How much do paid ads cost on Snapchat, TikTok, Google and Facebook?
There is no fixed price for an ad. Platforms sell impressions in an auction, and the price changes with the audience, the season, the number of advertisers and the quality of your ad
What platforms set is only the minimum budget: TikTok requires a daily campaign budget above $50 and an ad group budget above $20 (TikTok help center), Snapchat lets you start from $5 a day (source), and Google and Meta allow very small daily budgets. In the US market, the average cost per click on Google search ads was $5.42 in the LocaliQ 2026 report
Prices go up in seasons when advertisers crowd in, like Ramadan, Black Friday and back to school, because more people enter the auction. That is why the timing of your budget is part of its cost, as we saw in the Ramadan demand timing case study
A better question than "how much does an ad cost?" is "how much can I afford to pay for one order?" The same cost per click can be cheap for one business and a loss for another
What about influencer ads?
Same math: divide what you paid the influencer, plus the value of the discount their code gave, by the delivered orders from that code, then compare the result to your maximum cost per order. An influencer paid $1,500 who brought 60 delivered orders cost $25 an order, above Majed's limit of $16.91 no matter how many views the post got
How big should your digital marketing budget be?
Start from the goal, not from a fixed share of sales:
If Majed wants 150 orders a month at a target cost of $13.91, he needs about $2,087, as long as the cost of extra orders stays under his limit as he scales. If your budget is below the learning threshold, put it into one campaign on one platform instead of splitting it. To build the whole marketing budget from your sales goal, and know when to raise or cut it, read the marketing budget guide
Sales or ads: which comes first?
Sales come first. Not big sales, but proof that people buy from you at this price before you pay to bring in more of them
Ads are a loudspeaker that amplifies what already exists. If the offer is good, the price is right and the sales team replies fast, ads multiply sales. If any of that is broken, ads multiply the loss just as fast. This is the order we always work in:
- Product and price: is anything left from a sale before ads? If what is left is zero or close to it, no ad budget can fix that. We covered this in the pricing guide
- First customers without ads: from your network, direct outreach or your own page. The first 10 or 20 customers tell you whether the problem is real and the price acceptable
- Offer and page: a clear page, a visible price, an easy way to pay and a reason to buy now
- Measurement: you know where every order came from, how many were actually delivered and how many came from new customers. Without this, every number after it is a guess
- Sales capacity: in services and B2B software, is someone answering inquiries within an hour? An inquiry that waits a day goes cold
- Ads, then scale: a small budget that proves your cost per order is under your limit, then gradual increases while you watch the cost of the extra orders
When do ads come first?
When the ad is a research tool, not a growth tool: a small, fixed budget to test whether there is demand for an idea before you build it, or which message pulls people in. Here you are buying information, not sales, so judge it by what you learned, not by the return
Stores: is the order you bought worth its price?
Back to Majed. We calculate his ads step by step, from the number the platform gives him to the number that decides whether he should double the budget
How to calculate ROAS: what does a return of 4 mean for Majed?
ROAS (return on ad spend) is the sales that came from ads divided by what you spent on them. A return of 4 means every ad dollar came with $4 in sales
Majed spent $1,800 a month: $1,000 on Snapchat and $800 on TikTok. Snapchat said it brought 78 orders and TikTok said 64, so 142 orders. But Majed's own store recorded only 120 orders for the whole month, 108 were delivered, and the rest were refused at the door
How many orders really came in?
What each platform reports, compared to what the store recorded and what reached customers, in the same month
Each platform counts an order for itself if the customer saw its ad before buying, even if they also saw the other platform's ad or would have bought anyway. That is why the platforms' total is always bigger than reality
The way orders are credited to ads is called attribution (how a sale is credited to an ad). Each platform uses its own method and they do not agree, so the first rule is: the source of truth is your store or your shipping company, not the ads dashboard
With real numbers: 108 delivered orders at $50 are $5,400 in sales, divided by $1,800, a return of 3.0, not 3.94. This return calculated from all sales over all ad spend is called MER (marketing efficiency ratio), and it is more honest than any number inside a single platform
But even a return of 3 does not tell Majed whether he is making money, because he does not yet know what return he needs
Breakeven ROAS: the number to know before any ad
From the pricing guide we know the waterfall of Majed's bottle: the customer pays $50, which covers tax, the perfume, packaging, shipping, payment fees, returns and discounts. If we stop the waterfall before the ad step, every delivered order leaves $16.91
This is Majed's limit: if he pays less than $16.91 per order for ads he makes money, if he pays more he loses. Right now he pays $1,800 divided by 108, which is $16.67, so he keeps only 24 cents
We can write the same limit in the platform's language, as a return on ad spend. It is called breakeven ROAS:
For Majed: 50 divided by 16.91 is 2.96. So his return of 3.0 is barely above breakeven, and every ad dollar brings back less than a cent and a half in profit, before the platform subscription, software and photography
| Share of price left before ads | Breakeven ROAS | Example |
|---|---|---|
| 20% | 5.00 | Electronics, online grocery |
| 34% | 2.96 | Majed today |
| 50% | 2.00 | Private label apparel |
| 70% | 1.43 | Digital products, private label skincare |
That is why there is no single "good ROAS" for everyone. A return of 3 is excellent for a store that keeps 70% of the price and a loss for one that keeps 20%. Anyone who says "the market normal is 4" without knowing your costs is giving you a number that is not yours
A note on tax and returns
Some platforms count order value with tax and some without, depending on what your store sends. And every platform counts the order at the moment of purchase, before the customer refuses it at the door. So Majed's limit inside the platform is not $16.91 but $15.22, because 1 in 10 orders will not be delivered
Calculate your maximum cost per order
The numbers below are Majed's. Enter your product's numbers
The ladder: from simple math to the way pros calculate
The same ads from Majed, calculated six ways. Each one is more accurate than the one before, and each answers a different question
Return as the platform reports it
Sales the platform credits to itself ÷ what you spent there. Quick and useful to compare campaigns on the same platform, useless for judging profit
Real return (MER)
All sales actually delivered ÷ all ad spend. This is what happened in your bank account
Breakeven ROAS
Price ÷ what is left before ads. Anything below it means ads are losing money
Return that covers the whole month
With the same number of orders, what return covers ads plus $800 in fixed monthly costs? Maximum budget $1,027, or $9.51 per order
New customer cost
30 orders came from returning customers. Divide the ads by the 78 new customers only and each one cost $23.08, more than the $16.91 a first order leaves
Return on the extra dollar
When Majed doubled the budget, the extra $1,800 brought only 67 orders, a return of 1.87 on the increase, far below breakeven
Most stores stop at step one, and pros start at step three. Steps five and six are what separate someone who runs ads from someone who runs growth, and we explain them now
What happens when you double the budget?
When you raise the budget, the platform cannot find the same people again. It starts with those most ready to buy, then the more you spend, the more you pay to reach people who are less interested. So each extra order costs more than the one before it. This is called diminishing returns (each extra dollar buys less)
Majed raised the budget from $1,800 to $3,600 and delivered orders went from 108 to 175. The average cost per order became $20.52, which looks acceptable, but the extra orders alone cost $26.69 each, and every one of them lost about $10
The rule growth managers use: keep raising the budget while the cost of an extra order is below your maximum cost per order, and stop when they are equal, not when the average is
Monthly profit at every budget
Move the budget and watch orders, cost and profit change, in two cases: the order as it is today, and after fixing it
We assume orders grow with budget at a saturation factor of 0.7, so doubling the budget adds about 62% more orders. The second case: price $60, no discount codes and 5% returns, assuming the higher price cost Majed 1 in 10 buyers
Look at the red line: with the order as it is today, no budget makes Majed profitable. His best budget is about $576 and even there he loses $553 a month. The problem is not the ads, it is the order itself
After fixing the order, as we explained in the pricing guide, spending more starts to make sense: profit rises up to a budget of about $2,600 and then starts to fall. There is a best budget, and it is not "the biggest budget possible"
Majed did not need a bigger budget. He needed an order that leaves enough to pay for the ad
Where does cost per order come from? CPM, CTR, CPC and conversion rate
Cost per order is not a number you choose. It is the result of a chain: the ad is shown to some people, some of them click, some of those buy, and some orders come back. Every link has a name and a number:
- CPM (cost per thousand impressions): what you pay for your ad to show a thousand times. Majed pays $5
- CTR (clickthrough rate): out of every 100 who saw the ad, 1 clicked
- CPC (cost per click): comes from the two numbers above, $5 divided by 10 clicks, or $0.50
- Conversion rate (share of visitors who buy): out of every 30 who entered the store, 1 bought, or 3.33%
- CPA (cost per acquisition): $0.50 divided by 3.33% is $15 per order in the platform, and $16.67 per delivered order
This formula says something important: your store's conversion rate affects cost per order as much as the price of the ad. Improving the product page so that 4 in 100 buy instead of 3.33 cuts the cost per order from $16.67 to $13.89 without touching the ad itself
If one number improves by 20%
Majed's cost per delivered order. The dashed line is the maximum cost per order
Which of the four numbers is easiest to improve in your store? That is where you start, not where your ad agency wants to talk
Calculate your ad funnel
Enter numbers from your ads manager, and the maximum cost per order from the calculator above
Why does your competitor pay less than you for the same impression?
Ads are sold in an auction that runs every time someone opens the app or searches Google, but the auction does not simply go to the highest bidder. Google ranks ads by what it calls Ad Rank (the position score of an ad), a mix of your bid and the quality of your ad and page. The Meta auction combines your bid with its estimate of how likely the person is to engage with the ad and its quality
In practice: an ad people click more pays less to show, and a page people buy from more makes the platform show your ad more. Good creative is not a luxury, it is a discount on every impression
Why are Majed's results so unstable?
Meta's algorithm needs about 50 results a week per ad set to learn who buys. Before that it is in what it calls the learning phase (the period when delivery is still optimizing), where results swing and cost more
Majed has 120 orders a month spread across two platforms and 6 ad sets, about 5 orders a week per ad set. Not one ad set leaves learning. For one ad set to leave it he needs 50 orders at $15, or $750 a week for that ad set alone
The fix is not a bigger budget, it is less splitting: one main platform, fewer ad sets and more ads inside each. Google works the same way: target ROAS (bidding to a return goal) and target CPA (bidding to a cost per order goal) strategies work better the more results they learn from
New customer cost: a new customer is not an old one
Of the 108 orders, 30 came from customers who had bought from Majed before. Most of them saw a retargeting ad and would likely have bought without it. Spread the ads over all orders and the cost per order looks like $16.67. Spread them over new customers only, the real goal of ads, and it becomes $23.08
Does that mean Majed is losing money? He loses $6.16 on the first order from each new customer. But if 45 in 100 customers buy again within a year without a new ad, each customer leaves $24.52 a year, a little more than they cost
This is the question every store must answer with its own numbers: do I make money on the first order, or lose on it and make money when the customer comes back? Losing on the first order is not a mistake if you know how many customers come back and after how many months, and you have enough cash to wait. It is a big mistake if you assume they come back without measuring it
Order value and discounts: two levers most people ignore
Two bottles for $90
If Majed offers two bottles for $90, that order leaves about $35 before ads, because shipping, packaging and order fees are paid once. If 3 in 10 customers pick this offer, the average left per order rises from $16.91 to $22.34, so Majed can pay 32% more for ads per order. This is the effect of AOV (average order value)
20% off everything
A discount lifts the conversion rate so the campaign looks successful, but it eats what the order leaves. With 20% off every order, Majed keeps only $9.96 before ads, and his breakeven ROAS becomes 4.02 instead of 2.96. A sale campaign with a return of 3.5 looks better than usual and is actually losing money
What about a physical store?
Same logic, but measurement is harder because the customer does not click and then buy. Three practical ways: a unique code per campaign that the customer mentions at checkout, store visit estimates from the platforms treated as estimates rather than facts, and the most accurate: pause ads in one area or city for two weeks and compare its branches' sales to the branches where ads kept running
Service businesses: ads do not bring clients, they bring inquiries
In services there are many steps between the ad and the contract, and every step uses your team's time, and team time is the product
Sara and the $12 lead
Sara owns the design and marketing agency we met in the pricing guide. She spends $1,500 a month on Google Ads and pays about $2.5 per click, which brings 600 visitors, 30 of whom leave their details. So every lead, meaning an inquiry, costs her $50
Of the 30 inquiries only 12 are qualified: they have a suitable budget, the decision maker is the one who reached out, and the need is real. 9 attended a meeting, Sara sent 6 proposals, and two clients signed
Inquiries from ads
30Qualified
12Attended a meeting
9Received a proposal
6Signed
2Another month Sara tried Meta instant forms, which people fill in without leaving the app. The cost per lead dropped to just $12 and she got 125 leads on the same budget. She was happy, then her team spent the month calling: many numbers did not answer, many clicked by mistake, only 10 were qualified, and one client signed
To know which month was better, dividing ads by contracts is not enough. You must add your team's time: a quarter of an hour for the first call with each inquiry, an hour and a half per meeting with prep, and 3 hours per proposal. Sara's hour costs $27.34, as we calculated in the pricing guide
| Channel | Cost per lead | Cost per qualified | Client cost from ads | Sales time | Fully loaded client cost |
|---|---|---|---|---|---|
| Google Ads | $50 | $125 | $750 | $1,066 | $1,283 |
| Meta forms | $12 | $150 | $1,500 | $1,524 | $3,024 |
| $100 | $160 | $800 | $437 | $1,237 | |
| Referrals | $0 | $0 | $0 | $424 | $212 |
The cheapest lead produced the most expensive client, and the expensive LinkedIn lead produced a client costing about the same as Google. And look at referrals: they show up in no ads dashboard and are by far the cheapest channel. That is why every service business should ask each new client: how did you hear about us?
Customer acquisition cost (CAC) in services is ads plus sales time, divided by new contracts, not ads alone divided by inquiries
Customer lifetime value and the maximum price of a lead
Instead of asking "is $50 a lead expensive?", start from the client and work backward:
- What is a client worth? A first project of $3,000 at a 34.5% margin leaves $1,035. And 1 in 4 clients becomes a $2,500 monthly retainer at a 56% margin for 6 months on average, adding $2,100 per new client. Lifetime value (LTV) = $3,135
- How much can you pay to win one? The common rule is lifetime value of 3 times the cost, so $1,045 at most
- Subtract sales time: one client needs 15 inquiries and 19.5 sales hours, which is $533 of team time, leaving $512 for ads per client
- Step back one stage at a time: 2 in 9 meetings sign, so the most you can pay for a meeting is $113.7. 3 in 4 qualified leads attend, so the most for a qualified lead is $85.31
- Maximum cost per lead: 4 in 10 are qualified, so the maximum cost per lead is $34.12. Sara pays $50, 47% above the limit
The same math explains what happened with Meta forms: when only 8 in 100 were qualified, the maximum cost per lead was $6.83, so the $12 that looked cheap was almost double the limit
Maximum cost per lead for your business
The numbers below are Sara's
Three things that cut client cost more than ads do
1. Reply within an hour
A study published in Harvard Business Review sent an inquiry to 2,241 US companies: 37% replied within an hour, 23% never replied, and the average reply took 42 hours. In another part of the same research covering 1.25 million leads, companies that called within an hour were nearly 7 times as likely to qualify the lead as those that waited longer than an hour, and more than 60 times as likely as those that waited a day (study summary)
If Sara lifts meeting attendance from 75% to 90% just by replying fast, she gets more clients from the same ads without an extra dollar
2. Tell the platform who actually signed
Google and Meta do not know which inquiry became a contract, so they look for people similar to those who fill in the form, not those who sign. If you send them the outcome of each inquiry from your CRM, through offline conversion import (uploading sales results to Google) or enhanced conversions for leads (matching leads to their outcome) in Google, and the Conversions API (a direct connection to Meta), the ads start looking for qualified people
In Sara's case: if the cost per lead rises to $60 but qualified leads become 60% instead of 40%, and she replies fast, she gets 3 contracts instead of 2, her fully loaded client cost drops from $1,283 to $988, and lifetime value reaches 3.2 times the cost. This is exactly the kind of connection we build in our measurement service, with an example in the marketing attribution layer case study
3. Ask "how did you hear about us?"
One simple question in the contact form and on the first call reveals channels that show up on no dashboard: referrals, podcasts, content and word of mouth. Owners often discover that ads were taking credit for clients who came from somewhere else
Subscriptions and SaaS: you pay for the customer today and get it back in installments
The question here is not only "what does a customer cost?" but also "after how many months does the customer cover the cost, and how many are still around by then?"
Customer acquisition cost (CAC): does Khaled's clinic cost $300 or $600?
Khaled sells appointment booking software to clinics for $80 a month and keeps $64 after the cost of service. Every month he spends $6,000 to win customers: $3,000 on ads, $2,500 for a sales rep and $500 for free trials and onboarding, and 10 paying clinics come in
If Khaled compares clinic value to ad cost alone, he sees 5.3 times and thinks he is doing great. At full cost it is only 2.7 times, below the common threshold of 3. The gap is not an accounting detail: if he grows ads he will need another sales rep, so the full cost is the one that repeats
The real payback period
In the pricing guide we said a clinic needs about 10 months to cover its cost: 600 divided by 64. That math assumes all clinics stay, but 4 in 100 cancel every month. Once we count the clinics that cancel along the way, the average clinic only covers its cost in month 12
What a clinic has left so far compared to its cost
Per new clinic, in dollars, with and without cancellations
Churn (monthly cancellation rate) does not only change customer value, it delays payback. The gap grows the higher your customer cost
Ads bring trials, the product does the selling
$3,000 in ads at $2 per click gives 1,500 clicks, from which 90 free trials start. But only 36 clinics actually use the software in the first week and book their first appointment through it, and 10 of those pay
The moment the customer first experiences the value of the software is called activation, and it is the strongest number in a subscription funnel. If Khaled lifts activation from 40% to 55%, with a welcome call and help adding the first doctor, he gets 13.75 clinics instead of 10 from the same ads, and the full cost drops from $600 to $436
Fully loaded clinic cost at each activation rate
Same ads and same salaries
Maximum cost per click
If Khaled wants each clinic to cover its cost within 12 months, counting cancellations, the maximum full cost per clinic is $620. Ads are half of that today, so the most for ads per clinic is $310, per trial $34.43, and the maximum cost per click $2.07. Khaled pays $2, right at the limit, and any rise in cost per click without better activation pushes him over
When does your customer cover their cost?
The numbers below are Khaled's
Annual plan paid upfront
If Khaled offers a full year for the price of 10 months, the customer pays $800 today, more than the full clinic cost of $600. Payback happens on day one instead of month 12, and if 3 in 10 customers pick it, Khaled collects $2,960 in the first month from every 10 clinics instead of $800. This discount funds the ads, the opposite of free months that delay them
B2C: same logic, smaller numbers
In consumer apps the advertiser usually pays per install, and the numbers that rule everything are: how many installers start a trial, how many pay, and how many are still there after a month. Consumers cancel much faster than businesses, so the annual plan and first day activation matter more than the cost per install
Price and ads: whoever can pay more wins the auction
The most important ad decision is not made inside the ads manager. It is made when you set the price
In an auction, the advertiser who can afford to pay more for a customer wins the best placements and can keep going when others stop. And what decides how much you can pay? What the order leaves before ads, meaning price minus costs
If Majed raised the price from $50 to $60 and left everything else the same, what the order leaves before ads would rise from $16.91 to $25.06, 48% more, and his breakeven ROAS would fall from 2.96 to 2.39
But a higher price means some people will not buy, so the conversion rate drops and cost per order rises. The question: how far can it drop before raising the price becomes a mistake?
A $60 price: cost per order if fewer people buy
Cost per delivered order compared to the maximum cost per order at each price
Even if 25% fewer people buy, Majed keeps $2.84 per order after ads, more than 11 times what he keeps today
This is what economists call price elasticity (how much demand drops when price rises). You do not need theory to know yours, you need a test: two pages with two prices, or a new price for two weeks, then compare the conversion rate and what each order leaves, not just the number of orders
The company with the cheapest price in the market cannot outbid everyone on ads, and that is why it struggles to grow with ads
Ads with a growth mindset
An ads manager asks: how do I lower cost per order? A growth manager asks: how many orders came because of the ad that would not have come without it?
The order that would have come anyway
In 2012 eBay stopped its ads on its own brand name on Yahoo and Bing and kept them on Google for comparison. The result: 99.5% of the traffic that used to come from those ads came from regular search results at no cost. The researchers also found that search ads on generic keywords were losing money in the short term (published paper)
At Uber, the marketing team cut about two thirds of an app ads budget of around $150 million and saw no meaningful drop in new users, then discovered that part of the ads had been taking credit for installs that would have happened anyway (full story). In 2017 Procter & Gamble cut $200 million from digital ads and moved part of it to other channels, and its reach grew 10% (source)
Incrementality is the orders that came because of the ad and would not have come without it, and it is the only number worth paying for
How do you measure incrementality?
Pause and compare
Pause ads in one city or region for two or three weeks, keep them running in similar regions, then compare sales. The difference is the ad's effect. The simplest and most honest test for any business that sells in more than one region
Platform tests
Conversion Lift (measuring incremental conversions) in Google and a similar test in Meta: the platform hides the ad from a random group and compares their purchases with those who saw it
Brand is not a luxury
In 2021 Airbnb decided to cut ads that chase direct bookings and moved the money into brand building and PR. In 2022 it had its first full year of profit, with $8.4 billion in revenue, and its CEO said nearly 90% of traffic comes directly, without ads (Marketing Week)
The reason: ads that chase a purchase now harvest demand that already exists, while ads that build the name plant demand that comes later. If you always harvest and never plant, the harvest gets more expensive every month. Research by Les Binet and Peter Field for the IPA suggests that consumer brands put about 60% into brand building and 40% into direct response on average. The split varies by sector and company size, but the idea holds: do not put the whole budget into harvesting
How does a growth manager split the budget?
| Part | Share | Where it goes |
|---|---|---|
| Proven | 70% | Campaigns where the cost of an extra order is below your limit |
| Scaling | 20% | Audiences, platforms or offers that have started to prove themselves |
| Tests | 10% | New ideas, new messages, channels you have not tried |
Every test needs a clear hypothesis before it starts, one number to judge it by, and enough results before you decide, because 5 orders against 3 does not mean the first ad is better, it may be chance. And on today's platforms creative has become the targeting: the platform picks who sees the ad based on who engages with it, so the most important thing to test is the message, the image and the video, not audience details. And if you are looking for channels you have not tried outside Google, Meta and TikTok, the media buying guide covers where ads are sold in newsletters and niche sites and how to work out the maximum price for any placement
Every sector and its numbers
The logic is the same in every sector: know what the "order" leaves before ads, and know what it really costs you. But the "order" and the trap change:
| Sector | What ads buy | The number to judge by | Common trap |
|---|---|---|---|
| Ecommerce | An order | Cost per delivered order against what it leaves | Judging by platform ROAS, forgetting returns and discounts |
| Physical store, restaurant | A visit or delivery order | Sales in the area with and without ads | Crediting all sales to ads |
| Clinic, beauty center | A booking | Cost per patient who showed up, and how often they return | Counting bookings, not visits |
| Real estate | An inquiry | Cost per viewing, then cost per deal | Many cheap inquiries with no buying power |
| Education, courses | A signup | Cost per student who paid and finished | Free signups that never convert |
| B2B services | An inquiry | Fully loaded client cost including sales time | Judging by cost per lead |
| Subscriptions, software | A trial | Payback period with churn | Counting ads alone as customer cost |
Average cost per lead in Google search ads by industry
In dollars, from the LocaliQ 2026 report on US accounts
Numbers in your market differ, but the gaps between sectors are useful: lawyers pay more than 4 times what restaurants pay per lead because one client is worth more. The overall average is $66.69 with a conversion rate of 8.18% (report)
Before you raise the budget
- Do I know what an order leaves before ads, after tax, returns and discounts?
- Do I know my breakeven ROAS, and is the real return from my store or CRM above it?
- Am I counting delivered orders and signed contracts, not what the platform reports?
- How many of my results come from new customers, and what does a new customer cost on its own?
- What did the extra orders cost in my last budget increase, not the average?
- Do my campaigns leave the learning phase, or is the budget split across too many ad sets?
- Have I tried improving conversion rate, activation or reply speed before adding ad spend?
- In services and software: do I count sales time in customer cost?
- Do I have enough cash to wait until a customer pays back their cost?
- When did I last pause a campaign to see what happens without it?
ROAS, CPA, CPC and CPM explained
Each term links to a page that explains it at its source
| Term | Meaning | Formula |
|---|---|---|
| ROAS | Return on ad spend | Sales ÷ ad spend |
| Breakeven ROAS | The return where ads neither make nor lose money | Price ÷ what is left before ads |
| MER | Marketing efficiency ratio | All sales ÷ all ad spend |
| CPA | Cost per order or result | Ad spend ÷ orders |
| CPM | Cost per thousand impressions | Ad spend ÷ impressions × 1000 |
| CPC | Cost per click | Ad spend ÷ clicks |
| CTR | Clickthrough rate | Clicks ÷ impressions |
| Conversion rate | Share of visitors who buy | Orders ÷ clicks |
| CPL | Cost per lead | Ad spend ÷ inquiries |
| CAC | Customer acquisition cost | All sales and marketing cost ÷ new customers |
| nCAC | New customer cost on its own | Ad spend ÷ new customers only |
| LTV | Customer lifetime value | Monthly margin ÷ churn |
| CAC payback | Months to recover customer cost | Customer cost ÷ monthly margin |
| AOV | Average order value | Sales ÷ orders |
| Contribution margin | What a sale leaves after its variable costs | Price − every cost that repeats with the sale |
| Churn | Monthly cancellation rate | Cancellations ÷ customers at start of month |
| Incrementality | The extra effect of an ad | Results of those who saw it − results of those who did not |
| Attribution | Crediting a sale to an ad | Varies by model and platform |
| Diminishing returns | Each extra dollar buys less | Increase in spend ÷ increase in orders |
| Learning phase | The algorithm’s learning period | About 50 results a week per ad set |
| MMM | Marketing mix modeling | Statistical model of each channel’s effect on sales |
Free calculators
Three calculators on Google Sheets, one for each business type. They work as soon as you enter your numbers, and every tab has instructions. Click the button, then choose Make a copy, and you have your own private copy
Store ads calculator
For ecommerce and physical stores
- Maximum cost per order and breakeven ROAS
- Target ROAS for Google and Meta campaigns
- The funnel from impression to delivered order
- Budget scenarios and the best budget
- New customer cost
- Weekly log: platforms against store
Service business ads calculator
For agencies, firms and clinics
- Client value and maximum acquisition cost
- Backward funnel to maximum cost per lead
- Maximum cost per lead by lead quality
- Channels compared by fully loaded client cost
- Weekly log of reply speed
SaaS acquisition calculator
For software, apps and memberships
- Customer cost from ads and fully loaded
- Month by month payback with churn
- Maximum cost per trial and per click
- Effect of activation on customer cost
- Effect of the annual plan on cash
- Monthly log per channel
If you have not tried the pricing guide calculators yet, start there, because what a sale leaves before ads comes from there
FAQ
How much do paid ads cost on Snapchat, TikTok and Google?
There is no fixed price because ads are sold in an auction, and platforms only set a minimum budget, such as $50 a day for a TikTok campaign and $5 a day on Snapchat. The more important question is how much you can pay for one order before you lose money
What do CPA, CPC and CPM mean in marketing?
CPM is the cost of showing an ad a thousand times, CPC is the cost of one click, and CPA is the cost of one order or result. Cost per order equals CPM divided by a thousand times the clickthrough rate times the conversion rate
How big should a digital marketing budget be?
Multiply the number of orders or customers you want by your target cost per order, and make sure the budget covers about 50 results a week per ad set so the algorithm can learn. Otherwise put it all into one campaign
How do I calculate ad cost per order?
Divide what you spent on ads in a period by the orders delivered to customers in the same period, from your store or shipping company and not from the platform, then compare the result to what the order leaves before ads
How do I calculate ROAS, and what is a good ROAS?
Divide the sales that came from ads by what you spent on them. There is no single good number. A good ROAS is one above your breakeven ROAS, which is the selling price divided by what the order leaves before ads. A store that keeps a third of the price breaks even at 3, and one that keeps half breaks even at 2
What is the difference between ROAS and MER?
ROAS is the sales one platform credits to its own ads divided by what you spent there. MER is all your sales divided by all your ad spend. MER is more honest because platforms count some orders more than once
Why do platforms report more orders than my store recorded?
Each platform credits an order to itself if the customer saw or clicked its ad before buying, even if they saw another platform's ad or would have bought without ads. Platforms also count the order at purchase, before it is returned
Should I raise the budget when results are good?
Raise it gradually, about 20% at a time, and watch the cost of the extra orders rather than the average. Keep going while the cost of an extra order is below your maximum cost per order
What comes first, sales or ads?
Sales first, meaning proof that people buy at this price and that a sale leaves enough to pay for ads. Ads then amplify what is proven to work, unless the ad is a small test to check demand for an idea
How do I calculate the maximum cost per lead?
Start from client lifetime value and divide it by the multiple you want, usually 3. Subtract the cost of sales time per client, then multiply what is left by the close rate from meetings, the attendance rate from qualified leads, and the qualified share of leads
How do I calculate customer acquisition cost (CAC)?
Add everything you spent to win new customers in a period, including ads, sales salaries, free trials and tools, and divide it by the number of new customers in the same period
How do I know if ads brought sales or just took credit?
Pause ads in one area or city for two or three weeks, keep them running in similar areas, and compare sales. Or use the incrementality tests in Google and Meta
Ads do not fix a losing order. They make it lose faster
If you want us to review your ad numbers together: what an order leaves, your real limit, and where the budget goes without effect
Ehap Lil
I help companies reach clear, predictable financial growth by tying marketing effort to financial results, and by fixing the decisions that hold back returns or make sales swing
In this article
Table of contents
Read next
Marketing Budget: How Much to Spend, and Should You Spend More When Sales Drop?
Media Buying Beyond Google, Meta and TikTok: Where Else to Buy Ads
Pricing and Unit Economics: A Practical Guide for Stores, Services and SaaS
Collections, sales and financing are not separate islands
When should a startup hire a new employee?
