# SaaS growth agencies > How growth agencies differ from demand gen and performance shops, what they charge, the experiments they actually run, and when a fractional hire beats them. Source: https://saas-marketing.net/guides/saas-growth-agency/ Topic: SaaS Marketing Tools Type: guide Published: 2026-09-11 Last updated: 2026-09-11 Publisher: SaaS Marketing (saas-marketing.net) License: CC BY 4.0. Quote or republish with attribution and a link to https://saas-marketing.net/guides/saas-growth-agency/ ## Short answer A SaaS growth agency should produce four things: a growth model that links inputs to revenue, a sized and prioritised experiment backlog, instrumented funnels you can read, and a weekly decision cadence. Most shops calling themselves growth agencies produce media buying instead. Retainers typically run 8,000 to 30,000 dollars a month against a fractional growth lead at 6,000 to 15,000 and a fully loaded in house hire at 180,000 to 230,000 a year. ## Key takeaways - Growth agency is the least defined label in SaaS marketing, so judge every pitch by the work product, not the deck. - A real growth engagement produces a growth model, a sized backlog, instrumented funnels and a weekly decision cadence. - Typical growth retainers run 8,000 to 30,000 dollars monthly, against 6,000 to 15,000 for a fractional lead. - A 140,000 to 180,000 dollar in house growth hire costs 180,000 to 230,000 fully loaded, which changes the comparison. - Product led experimentation cannot be outsourced because it requires engineering access the agency will never get. - Outsource channels. Never outsource the growth model, because that is the asset you keep when the contract ends. --- "Growth agency" means nothing on its own. Three shops with identical homepages will sell you, respectively, a paid media team, a content operation with a conversion rate optimisation add-on, and two people who will actually rebuild how you think about your funnel. The price range across those three overlaps almost completely. So the only useful way to evaluate one is to ignore the positioning and ask what lands in your Google Drive at the end of month three. ## What a real growth engagement produces Four artifacts. If a proposal doesn't name all four, you're buying channel execution with a growth label on it. A growth model. A spreadsheet or dashboard that connects your inputs (traffic, signups, activation rate, trial conversion, expansion, churn) to revenue, with the sensitivities visible. It should let you answer "if we lift activation from 31 to 38 percent, what does that do to ARR in twelve months" without a meeting. Most companies don't have this and don't realise they don't. A sized experiment backlog. Not a list of ideas. Each item carries an expected impact, a confidence level, an effort estimate and the metric it moves. Prioritisation falls out of the sizing rather than out of whoever argued loudest. Instrumented funnels. Events defined, tracked and readable in whatever you run: Amplitude, Mixpanel, PostHog, or a warehouse. A depressing share of growth engagements spend their first six weeks discovering the client cannot measure activation at all. A weekly decision cadence. A recurring meeting with a fixed agenda where results get read and the next experiments get chosen. This sounds like process theatre and it's the thing that makes the other three compound. Ask what their last three engagements actually changed. A real growth practitioner answers with a specific funnel step and a before-and-after number. A media shop answers with a channel and a cost per lead. Neither answer is wrong, but only one of them is growth work, and you should know which you are buying. ## What it costs, against the two alternatives Here's the comparison almost nobody builds properly, because the in house number gets quoted as salary rather than fully loaded cost. A 150,000 dollar growth marketer costs roughly 195,000 to 210,000 once you add payroll taxes, benefits, equipment, software seats and the recruiting fee amortised over expected tenure. At the top of the band, a 180,000 dollar senior hire lands near 230,000. That's 16,000 to 19,000 dollars a month, which sits right inside agency retainer territory and changes the conversation entirely. The hidden cost of the in house route is hiring time. Three to six months from opening the requisition to a productive hire is normal for a senior growth role, and a failed hire costs you nine months and the recruiting fee. Agencies start in four to eight weeks and can be terminated in ninety days. That optionality is worth a real premium when you're uncertain about the work. The hidden cost of the agency route is context. An agency never knows your customers as well as your own team, and the gap shows up most in messaging work and in anything requiring product judgment. To run these numbers against your own salary bands and retainer quotes, the [agency vs in house cost calculator](/calculators/agency-vs-in-house-cost/) does the fully loaded arithmetic, and the market ranges are collected in [what SaaS marketing agencies charge](/research/saas-marketing-agency-pricing-benchmarks/). ## Why PLG experimentation cannot be outsourced This is the hard boundary, and it's where most growth agency engagements underdeliver against the pitch. Product led growth experiments live inside the product. Changing an onboarding flow, adding an activation prompt, restructuring an empty state, testing a paywall trigger: all of that requires engineering time, product roadmap negotiation and access to systems no agency will be granted. What actually happens is the agency writes recommendations, those recommendations enter the product backlog, and they sit behind three quarters of committed roadmap. Six months later the agency has produced a well-reasoned document and zero shipped experiments, and both parties are frustrated. Who ships the experiment? If the answer involves your engineering team, get a written commitment of engineering hours per sprint from your own CTO before the contract starts. An agency that has been promised two engineer-days a week and receives none will fail, and the post-mortem will blame the agency rather than the resourcing. Acquisition channels are different. Paid search, paid social, SEO, outbound, partnerships and lifecycle email all run in systems an agency can be given access to, on assets an agency can produce. That's genuinely outsourceable work and agencies are often better at it than a generalist in house hire. So the line is: outsource what happens outside the product, keep what happens inside it. And whichever side you're on, keep the growth model in house, because it's the only asset that survives the relationship ending. ## Contracting for work where the outcome is uncertain Growth experiments fail most of the time. That's the point of them. A contract structure that punishes failure produces an agency that only proposes safe tests, which is the worst possible outcome. Three structures and what they do to behaviour. Fixed monthly retainer. Predictable for both sides, and it lets the agency propose genuinely uncertain tests. The risk is drift into low effort activity once the initial build is done. Mitigate with a defined deliverable list and a quarterly review. Retainer plus performance bonus. Base fee covers the model, instrumentation and cadence; a bonus attaches to a specific funnel metric hitting a threshold. This works when the metric is genuinely within the agency's control. It corrupts fast when the metric depends on product releases or sales capacity. Pure performance. Almost always a bad idea for growth work. It pushes the agency toward channels with fast feedback loops and away from anything structural, and it produces arguments about attribution within two months. **What to put in the contract** The RFP stage matters more than people expect. A structured scorecard forces agencies to answer the deliverable questions rather than pitch case studies; our [agency RFP template and scorecard](/templates/marketing-agency-rfp-template/) is built for exactly this comparison. ## When each option is right Straight answers. Pick an agency when you need execution volume across channels you can't staff, when the work is bounded to two or three quarters, or when you need a capability like paid search management that doesn't justify a full time hire. Also pick one when you genuinely don't know what your growth constraints are, since a good diagnostic engagement is cheap relative to a wrong hire. Choose a fractional growth lead when you already have execution capacity, typically two or three marketers, but no one senior enough to set priorities. One or two days a week of real judgment fixes more than a full agency team in that situation, and it costs less. The wrong reason to pick fractional is expecting them to do the work; they won't have the hours. Pick an in house hire when the work is continuous, requires product access, and depends on knowing your customers deeply. Also when you're above roughly 8 million ARR, where the fully loaded cost is comfortably absorbed and the context compounds. The honest assessment of the three way choice, including what each option costs you in learning that walks out the door, sits in [agency, fractional or in house](/comparisons/agency-vs-fractional-vs-in-house/) and in the narrower [in house marketing team vs agency](/comparisons/agency-vs-fractional-vs-in-house/) comparison. **90 days** Maximum sensible lock-in on a growth agency contract before the first honest review ## The failure modes worth knowing about Three, and they account for most bad engagements. The pitch team disappears. The senior practitioner who impressed you in the sales process was in the sales process. Delivery goes to a coordinator two years into their career. This is why naming individuals in the contract matters. This scope quietly becomes reporting. Month one is diagnosis, month two is setup, months three through twelve are a weekly call where someone reads you numbers. If the experiment count per month is trending down and the deck is getting longer, that's the pattern. The engagement produces learning that leaves with the agency. Everything lives in their Notion, their spreadsheets, their heads. When the contract ends you have a dashboard and no understanding of how any of it was built. Force documentation into your systems from week one, not at handover. There's also a structural tradeoff nobody mentions in a pitch. An agency optimises for the duration of the contract, which is typically twelve months. Some of the highest return growth work, particularly SEO and lifecycle programs, pays back over eighteen to thirty six months. An agency has no incentive to start work that matures after they're gone, and a good one will tell you that directly. ## What to do before you sign anything Write down the four deliverables and send them to every agency on your list, before the first call. Ask each one which of their last ten experiments failed. Get the fully loaded in house comparison on the same page as the retainer quote, because the gap is smaller than most founders assume. And decide in advance which half of the work you're buying: if it's inside the product, you're probably hiring, not contracting. For the wider question of what belongs in your stack versus what belongs in someone else's, start with [the SaaS marketing stack](/saas-marketing-tools/), and if you're specifically evaluating shops rather than the model, our roundup of [SaaS Growth Agency](/guides/saas-growth-agency/) options and the broader guidance on [hiring a SaaS marketing agency](/guides/saas-marketing-agency/) cover the selection process. If the real gap turns out to be lifecycle and automation rather than strategy, [marketing automation for SaaS companies](/guides/marketing-automation-for-saas/) is the cheaper first move. ## Frequently asked questions ### What does a SaaS growth agency actually do? A genuine growth engagement builds a model connecting your acquisition and conversion inputs to revenue, produces a prioritised experiment backlog with expected impact sized for each item, instruments your funnel so results are readable, and runs a weekly cadence where decisions get made. Many agencies using the label instead run paid media, which is a channel service rather than a growth service. ### How much does a SaaS growth agency cost? Retainers commonly run 8,000 to 30,000 dollars a month, with most mid-market engagements between 12,000 and 20,000. Paid media management often sits lower but adds a percentage of ad spend. Project engagements such as a growth model build or a funnel audit typically run 15,000 to 45,000 as a one off. ### Should a SaaS company hire a growth agency or build in house? Hire an agency when you need channel execution you cannot staff, or when you need a specific capability for two quarters. Build in house when the work requires product access, deep customer knowledge or continuous iteration. The practical answer for most companies past 3 million ARR is a hybrid: in house owns the model and the product surface, an agency owns one or two channels. ### What is the difference between a growth agency and a demand gen agency? A demand gen agency is measured on pipeline generated from campaigns and channels, and works mostly above the funnel. A growth agency should be measured on conversion rate improvement across the whole funnel including activation and expansion, and works inside the product experience as well as outside it. In practice many demand gen shops have rebranded without changing their work. ### When does a fractional growth leader beat an agency? When you need judgment rather than hands. A fractional growth lead at one or two days a week can build the model, set the experiment priorities and mentor an existing junior team for roughly half an agency retainer. They are the wrong choice when you need actual execution volume, because one person at two days a week cannot ship campaigns. ### How should a growth agency contract be structured? Fixed fee for the model, instrumentation and cadence, because those deliverables are definable. Avoid pure performance pricing on experiments, since most experiments correctly fail and paying only for wins incentivises the agency to run safe tests. Include a three month break clause, a named team with minimum allocation, and a clause that all models and documentation are yours. ### What should you ask a growth agency before signing? Ask to see a redacted growth model from another client, ask which of their last ten experiments failed and what they learned, ask who specifically works on your account and for how many hours, and ask what happens to the model and documentation when the contract ends. The failure question separates real practitioners from account managers immediately.