Ask a software founder how they picked their Google Ads budget and the answer is usually a round number. Five thousand a month felt responsible. Ten thousand felt ambitious. Nobody worked backward from what a customer is allowed to cost.
The trouble is that a software budget doesn't buy customers directly. It buys clicks, and the clicks have to survive a sign-up form, a trial or demo, and a decision to pay before anything comes back. Each of those steps has a rate, the rates multiply, and the product of four ordinary-looking numbers can be a customer who costs more than they will ever pay you.
This chapter does that multiplication with real figures before any budget gets set. The click prices come from Keyword Planner's US data for software categories. The conversion rates come from two published datasets, one agency's and one industry survey's, each labeled for what it is. Your own numbers will be different. Run the same math on them.
A Software Customer Bought on Google Is a Chain of Four Numbers. The Weakest Link Sets the Price.
Every paying customer you get from Google Ads comes out of the same chain:
- Cost per click. What you pay each time someone clicks.
- Click-to-trial rate. The share of those clicks who sign up for a trial, open a free plan or book a demo.
- Trial-to-paid rate. The share of those who become paying customers. For a sales-led product, this is the close rate on demos.
- What a customer is worth. Monthly revenue, times your gross margin, times the months they stay.
The first three numbers set what a customer costs. The fourth sets what a customer is worth. The budget question is simply whether what a customer is worth can carry what they cost.
Divide the click price by both conversion rates and you get the cost of one paying customer in ad spend alone, before any salary, tool or onboarding cost. That number is the one most software founders have never written down.
One Agency's Paid-Traffic Data Puts a Paying Customer at About 81 Clicks. At $40 a Click, That's $3,240.
First Page Sage, a marketing agency, publishes conversion rates from its own 86 software clients, split by traffic source. For paid traffic to an opt-in free trial (no card required), it reports that 7.1% of visitors start a trial and 17.4% of trials convert to paid. The agency is clear that its conversion counts "even a single paid month," so a customer who pays once and leaves still counts.
Multiply the two rates and about 1.2% of paid clicks turn into a paying customer, roughly one in 81. At $40 a click, that is about $3,240 in ad spend per paying customer.
Is $40 a fair click price? For the buying searches Chapter 2 said to start with, it sits below the top of the range: "crm for small business" tops out near $60 at the top of the page, "project management software for small business" near $58, and "payroll software for small business" at $200. Those ranges are estimates from past auctions, and your actual price will usually land near them.
One agency's clients are not the whole market, and the cost per customer moves fast with every rate. Change the trial-to-paid rate from 17.4% to 10% and the same $40 clicks cost about $5,600 per customer. That sensitivity is the real lesson: small changes in the middle of the chain swing the price of a customer by thousands.
In a Survey of 200 Software Products, 1,000 Visitors to a Free Trial Made 3.6 Paying Customers. Requiring a Card Nearly Tripled It.
Kyle Poyar's Growth Unhinged newsletter, working with ChartMogul and ProductLed, surveyed 200 B2B software products and published a typical funnel per 1,000 website visitors, counting customers who start paying within six months:
- Free trials overall (most need no card): 45 sign-ups, 3.6 paying customers.
- Card-required free trials: 35 sign-ups, 10.5 paying customers.
- Freemium: 90 sign-ups, 5 paying customers.
The report also found that sign-ups from paid marketing, paid search included, "convert at the lowest rates" of any source.
Those funnels mix traffic from every channel, so applying a click price to them is our own arithmetic, not the survey's. Still, the direction is hard to ignore. If 1,000 visits cost $40 each, the typical free trial produces a paying customer for about $11,000 in ad spend. The card-required trial does it for about $3,800. Freemium lands around $8,000.
Notice what the card does. A card-required trial gets fewer sign-ups and nearly three times the paying customers. Chapter 1 made the same point from the tracking side: a sign-up that costs the buyer nothing proves almost nothing. Here it shows up on the budget. Whether to ask for a card is a product decision, not an ad setting, but it changes what every click is worth.
Payback Turns the Price Tag Into a Yes or No. Bessemer's Rule Is Under 12 Months for Small-Business Software.
A $3,240 customer is cheap or expensive depending on how fast they pay it back. The test is the payback period: cost per customer divided by monthly revenue times gross margin. A customer who brings in $500 a month at an 80% margin returns $400 a month, so a $3,240 customer pays back in about eight months.
Two published yardsticks help you judge the answer:
- Bessemer Venture Partners, a venture firm, tells its cloud companies to aim for payback "<12 months" when selling to small businesses, "<18 months" for mid-market and "<24 months" for enterprise.
- Aleph and Benchmarkit, using actual results from 342 software companies, put the median payback at 16 months, 11 months for contracts under $5,000 a year and 22 months for contracts of $50,000 to $100,000.
These figures measure all sales and marketing spend, not just ads, so your Google Ads cost per customer should come in well under them. Use your own gross margin, your own monthly revenue per customer and your own churn. The yardstick is the payback period your business can survive.
A $15-a-Seat Product Can't Pay $40 for a Category Click. A $5,000 Contract Can.
Put two products through the same chain. Both pay about $3,240 in ads per customer.
- The low-price product. One CRM in our sample publishes a price of $15 per user per month. Say a typical new account has three users: $45 a month. At an 80% margin, that customer takes about 90 months, seven and a half years, to pay back the ad spend. Even with no costs at all, it would take six years.
- The higher-value product. A product sold at $5,000 a year brings in about $417 a month. At an 80% margin, that customer pays back in about ten months, inside Bessemer's small-business target.
That doesn't mean low-price software can't use Google Ads. It means it can't buy the same clicks. Its options are the cheaper parts of the auction:
- Its own brand name, which Chapter 12 covers. In one agency's published book of software accounts, brand clicks averaged about $3 against about $14 for everything else.
- Long-tail buying searches that name a niche, a job or an integration, which usually cost less than the category head.
- A card-required trial or a paid starter plan, which raises the share of sign-ups who pay.
- Leading the ad with the plan that pays back, such as a team or annual plan, rather than the cheapest seat.
A Demo Booked in Minutes and a Customer Who Stays Move the Math More Than a Lower Bid.
Most founders try to fix a bad chain by cutting the bid. Usually the bigger levers are further down.
Speed at the handoff. Ten of the eleven sites in our sample that offer a demo let the buyer book a meeting time right away, without waiting for a callback. A buyer who requests a demo is shopping now, and every hour they wait is an hour to book someone else's. Every point you add to the demo-to-close rate cuts the cost per customer as surely as a lower bid does.
Customers who stay and grow. Lifetime value isn't the first payment. High Alpha's survey of software companies found net revenue retention above 100% in every contract-size band it reported from $10,000 a year up, meaning the customers who stay spend more over time than the ones who leave take away. Expansion and low churn can turn a customer who barely pays back into one worth several times their cost.
What customers say about you. Reviews and referrals don't show up in Google Ads, but they change the math. A customer who recommends you to two peers is worth more than their own subscription.
Pages built for the search. The click-to-trial rate is the link your landing page controls. A visitor who searched for a payroll feature and lands on a page about that feature is more likely to start a trial than one dropped on the home page. Chapter 4 picks up from there.
Next: A Software Home Page Answers Every Search a Little and None of Them Well.
Once you know what a customer can cost, the next job is lowering the cost of the clicks that bring them. Chapter 4 covers how Google ranks ads and why the page behind the ad matters as much as the bid.
Every chapter, from tracking to brand campaigns, competitor campaigns and the rules for free trials, is on the guide's home page. The rates in this worksheet are only as good as your tracking, so Chapter 1 comes first.
On the software accounts we take on for Google Ads management, the budget gets set after this worksheet, never before it: four numbers in, one payback period out. Run it on your own product before the next budget meeting.
