Paying for Clicks, or Paying for Customers

Every marketplace founder can tell you what CAC — customer acquisition cost — stands for. Far fewer can tell you what theirs is before they've spent the money finding out.

That's not carelessness. It's the nature of paid media. You commit budget, run the campaign, and the number reveals itself afterwards. Sometimes it's fine. Often it's a good deal worse than the spreadsheet said it would be, and by then the money is gone.

It's worth being precise about why that happens — and then putting lifetime value next to CAC, because CAC alone still isn't enough to say whether a channel works.

Keep the numbers simple

Throughout this piece we'll use one marketplace example:

• Average transaction: $100
• Take rate: 10%
• Marketplace revenue per booking: $10

That $10 is what you earn. The other $90 belongs to the seller. Every acquisition decision should be judged against your $10 — not against the $100.

What you're actually buying with paid ads

When you run ads on Google or Meta, you are buying attempts. Clicks, impressions, attention. You pay for all of them, whether or not anyone ever books.

Say you're paying $1 a click, and 2% of clicks turn into a completed booking. For a marketplace, 2% would be a strong result — you're asking someone to sign up, browse, trust a stranger and pay, all in one session.

At $1 a click and a 2% booking rate, you need about 50 clicks per customer — so CAC is $50.

Look at where that money went. One dollar bought the click from the person who actually booked. The other $49 bought clicks from people who didn't. That's not waste in the sense of a badly-run campaign — that is the campaign. You cannot buy only the converting click, because nobody knows which one it is until afterwards.

Put that against the $10 you earned on the first booking. You spent $50 to get $10 back. The standard answer is lifetime value: you didn't buy one booking, you bought a customer who'll book again. That's often true — but only if the LTV maths actually closes the gap.

LTV: the other half of the equation

CAC tells you what a customer costs. Lifetime value (LTV) tells you what they're worth. The decision lives in the relationship between them.

A simple formula, using marketplace revenue:

LTV = $10 marketplace revenue × number of bookings over the customer's life

Say a customer books six times before they leave. LTV is $60.

Two ratios then decide whether a channel is healthy:

• LTV:CAC — LTV divided by CAC. A common healthy target is 3:1: three dollars of lifetime revenue for every dollar spent acquiring. Below 1:1 you're losing money on the customer. Above 3:1 you have room to grow.
• CAC payback — how many bookings until the customer has repaid what you spent acquiring them. LTV:CAC asks whether the maths works eventually. Payback asks whether you can fund the wait.

Against a $60 LTV, a $50 paid-media CAC is an LTV:CAC of 1.2:1. The customer eventually pays back — but only just, and only if they really do complete all six bookings. There's almost nothing left to fund the rest of the business.

(For a deeper pass on contribution margin, CAC mistakes and liquidity, see Marketplace Metrics, Explained.)

What each channel is actually for

Before comparing the maths, it's worth saying this clearly: these channels are not interchangeable. Each one does a different job. Treating them as competing ways to buy the same click is how founders end up disappointed.

• Google search ads — best when someone already has intent. They typed "hire a photographer in Manchester" or "rent a campervan Cornwall". You're capturing demand that already exists. Great for high-intent categories; expensive when the keywords are competitive; weak when nobody is searching for you yet.
• Meta ads (Facebook / Instagram) — best for creating demand, not just catching it. You can reach people who weren't looking, test creative quickly, and retarget visitors who almost booked. Strong for awareness and consideration; CAC often looks higher because you're interrupting a feed rather than answering a search.
• AI ads and answer surfaces — platforms like ChatGPT are starting to sell placements and product recommendations inside conversational answers. Still early, but the job is clear: show up when someone asks an AI which marketplace, tool or service to use. Useful for category discovery; not yet a mature volume channel for most operators.
• Influencer marketing — best for trust and reach inside a niche. A creator lends their audience and credibility. That can drive bookings directly, but it also builds brand, content and social proof that paid ads alone struggle to buy. Influencer works especially well when the creator's audience matches your category tightly — and even better when you can measure outcomes rather than just paying for a post.
• Affiliate and referral programmes — best for performance-based growth once you have customers (or partners) who will recommend you. You pay when a booking completes. Ideal for compounding growth, lowering blended CAC, and turning happy users into a channel — not for inventing awareness from zero.

A healthy acquisition mix usually uses more than one of these, in sequence: paid or influencer to seed awareness and the first cohorts, then affiliate and referral to grow more efficiently on top.

The question nobody asks

Here's what I'd put to any founder weighing up acquisition spend: what if you knew your CAC before you spent anything?

Not modelled. Not forecast from someone else's benchmarks. Set, deliberately, by you, in advance.

That's what a referral or affiliate programme gives you. You decide what a new customer is worth — a flat $10, or 10% of marketplace revenue — and that becomes your CAC. Not approximately. Exactly.

And you pay when the transaction completes. Not when someone clicks, browses, or abandons signup. When they actually become a customer and money changes hands. The people who looked and didn't buy cost you nothing.

Same customer, different channel maths

Hold the marketplace constant — $100 booking, $10 revenue, $60 LTV — and change only how you acquire the customer.

• Paid search / paid social. CAC ≈ $50 from the example above. LTV:CAC ≈ 1.2:1. Payback takes about five bookings. The channel can work, but the margin for error is thin on a 10% take rate.
• Influencer. Not a write-off — a different bet. A $2,000 package that drives 40 completed first bookings is a $50 CAC, same ballpark as paid social on direct response alone. Where influencer often wins is everything around that number: content you can reuse, trust you couldn't buy in an ad auction, and an audience that already cares about the category. Structure more partnerships on performance (affiliate-style commissions or bonuses for tracked bookings) and the CAC can look a lot healthier still.
• Affiliate / referral with a fixed reward. Pay $10 when a referred customer completes their first booking. CAC = $10. LTV:CAC = 6:1. Payback lands on the first booking.
• Affiliate with a revenue share. Pay 10% of marketplace revenue on referred bookings for a year. On six bookings at $10 each, you pay $6. Effective CAC is lower, and it scales with the value of the customer rather than with an ad auction.

Same product. Same $10 per booking. Same retention. Change how you buy the customer, and the ratio moves from tight to healthy — especially once affiliate and influencer are set up to pay on results.

Two habits keep this honest. First, judge LTV on your $10 of marketplace revenue, not the $100 transaction — using GMV will make almost any CAC look fine until cash runs out. Second, calculate CAC per channel. A blended figure that mixes cheap referral with expensive paid will hide which channels are actually working.

There's a second-order effect too. A referred or creator-driven customer often arrives with trust already attached. They tend to convert more readily and stick around longer — so LTV can improve as well as CAC, which widens the gap further.

Why this matters more the thinner your margins

If you're running 40% margins, you can absorb a bad quarter of ad spend and learn from it.

At a 10% take rate, you can't treat acquisition casually. When each booking earns you $10, a channel that costs $50 a customer needs those customers to come back — repeatedly — before the maths works. You're betting on retention you may not have proven yet.

A cost you set yourself — a fixed reward or a share of revenue — behaves differently. It can't run away past the economics you designed. If the channel produces nothing, you pay nothing. If it produces a lot, you paid a known slice of money that was already in the door.

That's why affiliate and referral tend to become the efficient core of marketplace growth, while paid and influencer play the jobs they're actually good at: intent, reach and trust.

What referrals won't do

I'd be doing you a disservice if I left it there, because there are real limits.

Referrals don't scale on command. You can triple ad spend this afternoon and see traffic tomorrow. You cannot triple your advocates overnight. The channel grows with your customer base.

Paid media and influencers buy reach and attention that referrals simply cannot create from nothing. If nobody knows your marketplace exists, word of mouth has nowhere to travel. Most marketplaces need both: paid and creator channels to build the first cohorts and keep demand flowing, then affiliate and referral to compound growth at a CAC you chose.

And those programmes need to be run, not just switched on. Rewards have to be worth having, the ask has to land at the right moment, and you need tracking that shows what's actually converting.

But the underlying trade is a good one. Paid media asks you to spend money to find out what a customer costs. Affiliate and referral let you decide what a customer is worth, check that number against LTV, and only pay when you get one.