B2B SaaS marketing agency, judged on pipeline

What is the number that decides everything?

CAC payback: how many months of revenue it takes to recover what you spent acquiring a customer. It is the one figure that tells you whether growth is a machine or an expense, and published benchmarks suggest it has been getting worse, with one 2026 report putting the median for companies between 5 and 50 million dollars of annual recurring revenue at around 18 months, up from roughly 12 to 15 in 2023.

Another survey, of several hundred companies, reported a median closer to 16 months. The exact number matters less than the direction and the fact that your own number is knowable while an industry median is not actionable.

The trap in this metric is the average. One analysis makes the point sharply: a blended CAC payback of 14 months routinely hides a situation where 70 percent of customers paid back in 9 months and the other 30 percent never will. The blended figure looks mediocre and manageable. The reality is that most of the business is excellent and a specific segment is destroying value, and those need completely different responses.

The practical instruction is to never look at payback blended. Segment it by acquisition channel, by company size, and by the plan the customer bought. The segment that never pays back is usually identifiable within a quarter, and the usual correct answer is to stop acquiring it rather than to improve the marketing aimed at it.

Why is your attribution lying to you?

Because the default model credits whatever a buyer touched last, which is almost never what actually convinced them. The brand search that closed the loop gets the credit, the podcast and the comparison article that started it get none, and you cut the channel that was doing the work.

The fix is unglamorous and it is mostly plumbing rather than analytics. Join your marketing data to closed won revenue in the CRM, so you can ask what a customer cost rather than what a lead cost. Add a self reported field on the demand form and read it, because for all its imprecision it captures the dark social activity your analytics cannot see. And accept a tolerance: in B2B software, attribution that is directionally right and understood by everyone beats a precise model nobody trusts.

This is also where being an engineering firm changes what we can offer. Joining CRM data to marketing data is an integration problem, and we build integrations. Most marketing agencies have to ask your developers for this and then wait.

Why did more leads make your pipeline worse?

Because lead volume and lead quality move in opposite directions when you optimise for cost per lead. Lower the cost per click, broaden the targeting, and you get more leads at a better headline price while the rate at which those leads become qualified opportunities collapses.

This is the most common way a paid programme looks successful on a dashboard and is destroying money in reality. The marketing report shows leads up and cost per lead down. Sales says the leads are terrible. Both are telling the truth.

The remedy is to stop optimising for the lead and start optimising for the thing after it. Pick the furthest down funnel event you can measure with reasonable volume, usually a qualified opportunity rather than a closed deal, and judge channels on cost per that. Volume will fall and the finance conversation will improve.

There is a related trap with traffic. Placements that drive visitors but no pipeline, a podcast appearance, a press mention, a viral post, feel like progress and frequently produce nothing measurable. They can still be worth doing, but they should be budgeted as brand rather than counted as demand, and the distinction has to be explicit or every quarter ends in an argument about what worked.

Why is it dangerous to cut a channel quickly?

Because B2B software buying cycles commonly run five to nine months, so a channel you switch off today stops producing revenue six months from now, long after everyone has forgotten the decision. The lag means channel cuts are evaluated against the wrong quarter.

This is the single most expensive dynamic in SaaS marketing and almost nobody plans around it. A new executive arrives, looks at a content programme that is not producing attributable pipeline this month, and cuts it. Two quarters later inbound pipeline falls and the cause is invisible, because the thing that caused it stopped half a year ago.

Three habits that protect against it. Give slow channels a stated horizon before you start, written down, so cutting early is a decision rather than a reflex. Measure leading indicators for those channels, rankings, branded search volume, demo requests that cite content, rather than only closed revenue. And when you do cut, cut deliberately and record the date, so when pipeline moves two quarters later somebody can connect the two.

Worth noting alongside this: one 2026 report describes paid acquisition falling as a share of the mix, from around a third in 2023 to roughly a quarter, with organic and content growing. If that direction is right, the channels with the longest feedback lag are the ones taking more of the budget, which makes the discipline above more important rather than less.

Which channels work at which stage?

Channel choice depends on how many customers you have, not on how much budget you have. Before product market fit, founder conversations. In the first fifty customers, founder outbound plus one narrow channel. After repeatability, the compounding channels earn their place. Agencies that recommend the same mix regardless of stage are selling a template.

Two errors recur often enough to name. The first is mistaking an audience for a market: people who read your posts and congratulate your launch are not buyers, and a large following that does not convert is a very expensive lesson. The second is treating a launch as a strategy. A launch produces a spike. If you cannot describe how customer one hundred and one arrives after launch week, you planned an event rather than a go to market motion.

A third, specific to founders who sell before hiring: close the first twenty to fifty yourself. Published advice on this is consistent, and the reasoning is that a sales hire can refine a playbook but cannot invent one. Hiring sales to discover your messaging is the most expensive form of customer research available.

What does SEO look like for a SaaS company?

Pages that answer what buyers type, not what you want to say, plus the technical quality to be eligible. For software that means comparison pages, alternatives pages, integration pages and problem led content, with the product pages pointed at the commercial terms. It compounds slowly and it is the channel most often cut too early.

The content that produces pipeline in B2B software is not thought leadership. It is specific and unglamorous.

Comparison and alternatives pages. Someone searching your competitor plus alternatives is further down the funnel than almost any other visitor. These pages are uncomfortable to write and they convert.

Integration pages. One page per tool you connect to. Buyers search for their existing stack, and this is also the pattern that earns links naturally.

Problem led content. The question your buyer types before they know a product category exists. Lower intent, longer payback, and it is what makes you the obvious choice when they are ready.

Free tools. A calculator, an audit, a scorecard. Expensive to build and cheap to rank, because almost nobody does it. This is the part most agencies cannot deliver, since it is software rather than content, and it is the clearest example of why we built our own free tools rather than writing another blog post about them.

What we would steer you away from: publishing volume for its own sake, keyword stuffed pages aimed at terms with no buying intent, and the belief that ranking equals revenue. Traffic that cannot be tied to pipeline is a cost centre with good reporting.

What about paid acquisition for SaaS?

Paid is for capturing intent that already exists and for compressing the feedback loop, not for creating demand cheaply. Run it on terms where the buyer has a problem they can name, measure it on cost per qualified opportunity rather than cost per lead, and expect the economics to be tighter than consumer marketing because the terms are expensive.

The arithmetic in B2B software is unforgiving. Clicks on commercial software terms routinely cost tens of dollars, which means a modest conversion rate and a long sales cycle can turn into a payback period that never arrives. The response is not to find cheaper clicks, which usually means less qualified ones, but to improve what happens after the click.

Three things that move paid performance more than bidding does. A landing page written for the specific search rather than a generic product page. A conversion offer appropriate to the stage, because asking for a demo from someone comparing options is asking too early. And negative keywords maintained properly, since a large share of wasted spend in software accounts goes to searches that were never going to buy.

Where paid genuinely earns its place beyond pipeline: it is the fastest way to test messaging. A week of paid traffic across three value propositions tells you something that would take a quarter to learn from content, and that finding then improves every other channel.

What makes Voliom different from a normal SaaS marketing agency?

We build the product as well as the pipeline. When a page needs a calculator to convert, when attribution needs your CRM joined to your analytics, or when the fastest growth lever is a free tool that ranks, we ship it rather than writing a brief and waiting for your engineers.

Most marketing agencies hit the same wall. The strategy says you need an interactive assessment, a comparison tool, or product analytics wired into the funnel, and then it sits in your engineering backlog behind the roadmap. Months pass. The strategy was right and nothing happened.

Three things that change when the same team does both.

Free tools instead of more blog posts. Our own site runs an MVP cost estimator, a go to market scorecard, a website audit, a security scanner and an enrichment tool. Each one is software, each one ranks, and each one captures intent that an article cannot. We build these for clients because we build them for ourselves.

Attribution that actually gets built. Joining closed won CRM records to first touch is an integration, not an analytics setting. We do integrations, so the measurement conversation ends in a working system rather than a recommendation.

Programmatic pages done properly. Generating many targeted pages from structured data works when the data is real and each page is genuinely useful. It fails, visibly, when it is spun. The difference is engineering, and we have done both the building and the recovering.

The honest trade: a specialist B2B SaaS marketing agency with ten years in one vertical will have sharper pattern recognition in that vertical than we do. If you need deep category expertise and nothing built, hire them. If your plan keeps stalling on engineering capacity, that is the problem we exist to remove.

How do you get found when buyers ask an AI instead of Google?

By being the source that answers the question directly. Assistants and AI summaries lift passages that state an answer in the first line, cite something checkable, and are structured enough to quote in isolation. The same writing that earns a featured snippet is what gets quoted by a model.

This is changing how discovery works in B2B software, and the agencies describing it as a new discipline with a new acronym are mostly describing good writing with a fee attached. What actually seems to matter is unglamorous.

Answer first, then explain. A passage that opens with the answer survives being quoted without its heading. One that opens with context does not, because the quoted fragment says nothing.

Be specific enough to be citable. A number with a source attached is quotable. A claim that something is important is not. This is also why vague marketing copy performs badly here: there is nothing in it to lift.

Cover the comparison questions. Buyers asking an assistant which tool to use get an answer assembled from whoever wrote clearly about the comparison. If your category has no honest comparison content, somebody else's framing becomes the default.

Keep the facts current and dated. Models and search both discount content that looks stale, and in software a page describing last year's pricing or a deprecated integration is worse than no page.

The uncomfortable part for marketers: this rewards content that answers the question completely, including when the honest answer sends the reader elsewhere. Pages written to withhold the answer until a form is filled are exactly the pages that do not get cited.

What should the first ninety days look like?

Measurement in weeks one and two, one channel in weeks three to eight, and a decision with evidence by week twelve. Any plan that begins with publishing before anyone has joined marketing data to closed revenue is a plan to be busy for a quarter and uncertain at the end of it.

1. 01

   Fix what you can measure

   Weeks 1 to 2

   Join marketing data to closed won records, add a self reported source field to the demand form, and agree which down funnel event channels will be judged on. Unglamorous and it determines whether the next ten weeks teach you anything.

2. 02

   Listen to customers

   Weeks 1 to 3

   Recorded sales calls, churn interviews, support tickets. Positioning written without this is guesswork dressed as strategy, and the language buyers use is almost never the language the company uses internally.

3. 03

   Run one channel properly

   Weeks 3 to 8

   One, chosen for your stage, run well enough to produce a real signal. Running four badly is the most common way a first quarter produces no usable information.

4. 04

   Decide with evidence

   Weeks 9 to 12

   Cost per qualified opportunity by channel, payback segmented rather than blended, and an explicit decision to continue, change or stop. Written down, with the date, so the next person can see why.

What we would not do in the first ninety days: a brand refresh, a full website rebuild, or a content calendar stretching out six months. All three feel like progress and none of them produce information. They are reasonable investments once you know which message converts, and expensive guesses before that.

What belongs in a monthly report?

Four things: cost per qualified opportunity by channel, payback segmented rather than blended, what changed and why, and what we got wrong. A report that only contains numbers that went up is not a report, it is an invoice with charts.

Agency reporting has a bad reputation it mostly deserves. Impressions, reach and engagement all rise reliably with spend and tell you nothing about whether the money worked. Here is what we think a useful monthly report contains.

Cost per qualified opportunity, by channel. Not cost per lead. The number that connects to revenue, with the definition of qualified agreed in advance and unchanged, because a moving definition makes every trend meaningless.

Payback by segment. By channel, by company size, by plan. The blended figure goes in for continuity and the segments are what you discuss.

Leading indicators for slow channels. Rankings, branded search volume, demo requests citing specific content. These are how you judge a compounding channel before revenue arrives, and they are the protection against cutting it early.

What we got wrong. A test that failed, an assumption that did not hold, a channel we recommended that is not working. An agency that never reports a miss is either not testing anything or not telling you about it, and both are worse than the miss.

One practical note on cadence. Monthly is right for reporting and wrong for decisions. In a market with a six month buying cycle, judging a compounding channel on a monthly number produces exactly the premature cuts described earlier on this page. Report monthly, decide quarterly.

B2B marketing agency, growth marketing, branding: which do you need?

A B2B marketing agency runs the function broadly, a growth marketing agency optimises an existing motion against a number, and a branding agency decides what you stand for and how you look. Companies usually hire the wrong one because the symptom they feel is not the problem they have.

The categories overlap enough in sales conversations to be unhelpful, so here is the distinction that actually matters: what has to be true before each one can work.

The most common mismatch we see: a company with a positioning problem hires a growth marketing agency, because the symptom is weak pipeline and growth sounds like the answer to that. The agency optimises campaigns against a message that was never going to work, both sides get frustrated, and the real problem is untouched six months later. The reverse happens too, where a company with a working motion and no capacity hires a branding firm and receives a beautiful document.

Our honest answer about ourselves: we are closest to the first row, with engineering attached, which matters when the plan requires building something. If what you need is the second row and you already have a working channel, say so, because the engagement looks different and costs less. If it is the third, the marketing audit is the cheapest way to find out whether positioning is really the constraint, and we publish the whole framework so you can run it without us.

When should you not hire a SaaS marketing agency?

Four situations, and we check all of them on the first call: you do not have product market fit, nobody on your side can answer questions within a day, you need pipeline next month but have a nine month cycle, or the real problem is sales rather than marketing.

The last one is worth dwelling on because it is the most commonly misdiagnosed. If leads arrive and then stall, adding leads makes the problem larger and more expensive. The diagnostic is simple: look at the conversion rate from qualified opportunity to closed won. If that number is poor and consistent, marketing is not your constraint, and an agency that takes the work anyway is taking money to make your problem worse.

How do you judge a SaaS marketing partner?

Six questions, none of them about tactics. What would you stop doing, how do you measure before you pitch channels, what happens in month two if nothing works, who actually does the work, can we speak to a client who left, and what do you need from us.

One more test that costs nothing. Describe a requirement slightly outside what they do and see whether they say no. An agency that can say "that is not us, you want a specialist in this" will also tell you when something is not working. One that says yes to everything will say yes to the thing they cannot do.

What should you have ready before the first call?

Four things, and none of them is a brief. Who your best twenty customers are and what they have in common, access to recorded sales calls, your current numbers even if they are unflattering, and one person who can answer a question within a day.

Your best twenty customers. Not your ideal customer profile document, the actual list. Patterns in who already buys and stays tell you more than any market research, and the ones who churned tell you as much again.

Recorded sales calls. Ten is plenty. This is the single most useful asset a client can hand over, because buyers describe their problem in words that almost never match the words on the website, and that gap is usually where the positioning work is.

Your numbers, including the bad ones. Current pipeline, win rate, average contract value, churn. If attribution is broken, say so rather than sending a tidy report that implies otherwise, because we will find out in week two and the tidy version costs everyone a fortnight.

One responsive person. Not a committee. Marketing work stalls on unanswered questions faster than on anything technical, and an agency left guessing will produce confident work aimed at a buyer who does not exist.

What you do not need: a creative brief, a channel preference, or a content calendar. Those are outputs of the first few weeks rather than inputs to them, and arriving with them fixed usually means paying for work to be redone once the sales calls contradict them.

What does a B2B SaaS marketing agency cost?

Retainers here are commonly quoted between three and fifteen thousand dollars a month, frequently plus a percentage of advertising spend, though that figure comes from an agency's own published comparison rather than independent data. We work differently: hourly against an agreed scope, with no percentage of your ad spend.

We do not take a percentage of media spend, and we would encourage caution about anyone who does. It creates an incentive to increase spending that is not aligned with your payback period, and it is the arrangement most likely to end with a large paid budget and an unexamined CAC.

Indicative until we have scoped the work. Most engagements start with a defined first project rather than an open ended retainer, because a retainer agreed before anyone understands the measurement is how both sides end up disappointed. You can read about the team, and the go-to-market strategy page covers the planning side if you are earlier than this.

• Go-to-market strategyWhen you are earlier than channels
• SaaS SEO agencySearch for software, after AI summaries
• SaaS PPC agencyPaid search, measured on pipeline
• Landing pages and CROSame traffic, more customers
• SaaS developmentWhen the product needs building too
• Brew, free enrichmentBuild the outbound list without credits
• About VoliomWho does the work