How to choose a SaaS marketing agency: match the kind of help to your stage

A SaaS marketing agency is an outside team paid to win customers for software sold by subscription (SaaS, or subscription software in plain terms). The label covers at least three different jobs, and quotes for different jobs can look alike on paper. A demand-generation shop runs paid media, landing pages and outbound sequences, and answers for pipeline. A content and search specialist works on organic discovery, the pages a buyer reaches through search engines and AI answers while weighing up products. A fractional marketing leader is a part-time senior hire, not an agency, yet is sold under the same label. Ember Tribe, a B2B marketing firm that sells strategy and execution work, says in its guide to fractional CMOs that such a leader is strategic leadership, not a full-time producer. A single firm may sell more than one of the three, so ask which job each line of a quote pays for.

Stage decides which kind of help you are buying

Before a sale repeats, the open question is who buys, why, and which message earns a reply, more than how many leads arrive. Data-Mania, a vendor in this market, puts the split simply: for strategic clarity, go with a fractional CMO, and with a validated strategy and a wish for help executing it, turn to an agency. It sells services of this kind, so read that as one vendor's rule, not a finding.

Our view is that a demand-generation shop hired before the buyer and the message are settled produces activity nobody can judge. A fractional leader hired with nobody to execute produces plans nobody runs. Deal size, growth rate and company scale also move every benchmark that follows. The stage bands in the table are an illustration of ours, and no published source sets them.

Number one: CAC payback, on a definition you have agreed

CAC is customer acquisition cost, and payback is how long a customer takes to repay it. Aleph, a finance-software vendor, defines CAC payback as the number of months of gross profit it takes to recover the sales and marketing cost of acquiring a customer. Its 2026 report with Benchmarkit found the median B2B SaaS company recovers that cost in 16 months, using full-year 2025 actuals across 342 SaaS and AI-native software companies. Its payback figures rest on the 198 participants that reported the metric. Top-quartile companies do it in 6 months or fewer, while the bottom quartile takes 24 months or more. Treat that as a survey of reporting companies, not of the whole market.

The median hides the stage effect. In the same data, companies selling contracts under $5K a year had an 11-month median, and those at $50K to $100K had 22 months. By growth rate, companies growing 21 to 30 per cent showed the longest median at 22 months, and those growing over 50 per cent showed 10. Aleph's own advice is to compare within your ACV band, where ACV means annual contract value.

Two details change the answer. Aleph says a payback period that ignores the cost of delivery flatters low-margin businesses. It also says to lag the spend, because the sales and marketing cost that closed this quarter's new ARR was largely spent in prior periods. So ask the agency which spend counts as acquisition cost, and whether your own retainer sits inside it.

Report it quarterly, by channel and by cohort, with the formula printed above the figure, because spend runs ahead of the sales it produces and a monthly figure mostly measures timing. One quarter's number matters less than the direction across three. Write the trigger into the contract: payback beyond your band for two quarters running obliges the agency to propose a change of plan in writing, with the channel it would cut.

Number two: marketing spend as a share of recurring revenue

SaaS Capital, a lender to software companies, runs an annual survey of private B2B SaaS businesses. Its 15th, completed in March 2026, drew more than 1,000 responses. The median share of annual recurring revenue spent on marketing was 8%, against 15% on selling costs. For a company with $3 million to $5 million in ARR, the medians were 8% on marketing and 12% on selling. The survey also found equity-backed companies spending 70% more on sales and 100% more on marketing than bootstrapped ones, so the funding route shifts the benchmark.

An illustration, not a budget: at $4 million of ARR, 8% is $320,000 a year. That median is a company's whole marketing spend, so a proposal whose fees, advertising and tools come near it would take all of it, and one well above it is asking you to spend like an equity-backed company. The question is what share of ARR the plan consumes in year one, and what it must deliver to be worth that.

Of the cost figures, this is the one to see every month, by line, with the agency's fee shown apart from media and tools. Agree the ceiling before the work starts and decide in advance what a pause looks like, so that crossing the ceiling triggers the pause you agreed rather than a debate about it.

Number three: free-to-paid conversion, for self-serve products

If people can start a trial without talking to sales, free-to-paid conversion matters. Kyle Poyar's study with ChartMogul and ProductLed was conducted in January 2026 and included 200 software products, 22% of them mostly B2C or a hybrid of B2C and B2B. It found a median free-to-paid conversion rate across all products of 8%. The survey defined it as the share of leads or free signups that become paying customers within six months. The spread is wide: among free-trial products, one in five sat below 2.5%, and free trials that require a credit card saw 30%, more than 5x the rate of those that do not. The report adds that a credit card requirement could blunt signups, hurting the overall number of paying customers.

An agency that quotes "the benchmark" without a window and a trial type is quoting nothing. Ask for conversion by signup cohort each month, split by card or no card and by acquisition channel, with any cohort that has not finished its window marked as an estimate. A finished cohort that converts below the one before it is the point to ask which channel changed. For a sales-led product this row does not apply, and the next one does.

Number four: the rate between a lead and a qualified opportunity

This is a labelled assumption, not a benchmark, and no figure is given. It moves faster than the other four, which makes it tempting to report early, and it means least until its terms are written down. A marketing-qualified lead, a sales-qualified lead and a demo that actually took place are three different events. A rate built from them changes with every choice about which one counts.

Before the first campaign, get a one-page definition of "qualified", signed off by whoever runs sales. Because this is the fastest of the five, it belongs in a short weekly note: leads sales accepted, demos held and opportunities created, with the previous quarter beside them. Read a sample of the rejected leads yourself. A falling accepted-lead rate in month two is the early warning. The agency should answer it in writing within an agreed number of working days, saying what it thinks happened, what it will change and which number will show whether the change worked. The definition itself is the one thing it may not change without your sign-off, because a rate fixed by redefining a lead has not been fixed.

Number five: net and gross revenue retention

Retention measures how much of last year's recurring revenue is still there. Gross retention counts only losses, and net retention also counts upgrades. SaaS Capital's 2026 benchmarks for bootstrapped companies with $3M to $20M in ARR show median net revenue retention of 103% and gross revenue retention of 91%. Top performers in the 90th percentile reach 117.9% on net retention.

An agency should report these. Judge the agency on acquisition-source effects within its control, not aggregate retention alone. Product quality, onboarding and support are run by your own team, and they move this number in ways no campaign can. What the agency does control is who it brings in. Ask for gross retention each quarter, split by acquisition source. If one source churns faster, you want to hear it from the agency first, together with how it would move spend away from that source. A company with a handful of customers has no retention number worth reading yet.

Who should hold off hiring, stage by stage

The stage table answers this too, read from the top. With no repeatable sale, do not hire a demand-generation shop or a content specialist. If closed deals come from the founder's network and nobody can say why the last five customers bought, neither has anything stable to improve. A fractional leader can help find the buyer and the message, but only if someone on the team can execute. Without that person, founder-led marketing with one tracked experiment a month builds the evidence any later hire will need.

With a repeatable sale but no payback figure you trust, hold off on a demand shop as well. Without a few quarters of closed deals and tagged sources, its payback figure is noise. A freelancer on one channel, such as search content, asks for less commitment while that data builds.

With gross retention below the median for your band, outside marketing help is unlikely to be the main fix. Acquisition source can play a part, which is why the retention section asks for it by source. But if customers leave at a similar rate from every source, new customers would be refilling a leaking bucket, and product and onboarding work comes first. And at any stage, if fee plus spend would take a much larger share of ARR than the medians above, the plan is outrunning the business. An in-house hire suits work needed every week for years, and the trade-offs are set out in hire an agency vs build in-house.

The Social Target has no SaaS case study, so there is no client result from this sector to show you. What it offers software companies is described on its marketing for SaaS and software companies page, which is a description of services, not proof.

If you want to test a quote against these five numbers for your stage, tell us about your business.