When SaaS Marketing Matures, Stop Optimizing Everything

Mature SaaS marketing teams face a specific tension that rarely gets named clearly: the better you get at optimization, the harder it becomes to justify experimentation. When your paid channels are humming, your email sequences are dialled in, and your attribution model is telling a coherent story, the case for trying something genuinely new looks weak against the case for squeezing another two percent out of what already works. That tension is not a sign of dysfunction. It is a sign that your team has grown up. How you resolve it is what separates teams that compound their advantage from teams that plateau.

I have watched this play out across a lot of different organisations. When I was scaling an agency from around 20 people to over 100, the teams that got stuck were almost always the ones that had found something that worked and then defended it too hard. The instinct to protect a working system is rational. But in marketing, a working system has a shelf life, and the teams that forgot that paid for it later.

Why Optimization Wins the Internal Argument Every Time

Optimization has a structural advantage in any internal budget conversation: it produces numbers faster. If you are running a test on a landing page variant or tightening bid strategies in paid search, you can show results in days or weeks. The feedback loop is tight, the attribution is clean, and the story you tell in the next performance review is easy to construct.

Experimentation is the opposite. A genuine experiment, the kind that tests a new channel, a new positioning angle, or a new audience segment, takes months to produce interpretable signal. The results are often ambiguous. The attribution is messy. And if it does not work, the person who championed it has to explain why they spent budget on something that produced nothing visible.

This is not a failure of nerve. It is a rational response to how most SaaS marketing teams are measured. If your quarterly review is built around MQL volume, CAC, and pipeline contribution, experimentation will always look like a liability. The incentive structures reward the optimization mindset, and over time, teams self-select toward it.

The problem is that optimization compounds diminishing returns. The first 20% efficiency gain on a paid search account is relatively easy. The next 20% takes three times the effort. At some point, you are fighting for fractions of a percent, and the opportunity cost of that effort, measured in experiments not run, becomes significant. How teams structure themselves around this challenge is one of the more consequential decisions a marketing leader makes in a scaling organisation.

What Experimentation Actually Means in a Mature Team

There is a version of “experimentation” that is really just optimization with extra steps. Running an A/B test on a subject line, testing two ad creatives against each other, or trialling a new bid strategy are all useful activities. But they are not experiments in the sense I mean here. They are refinements of an existing system.

Genuine experimentation in a mature SaaS marketing context means testing something that could change the shape of your demand generation model, not just improve the efficiency of the existing one. That might mean running a content-led motion in a segment you have historically approached through paid. It might mean testing a product-led growth mechanic alongside your sales-assisted funnel. It might mean investing in community or partnerships in a way that does not produce attributable pipeline for six months.

These are uncomfortable bets. They require a different kind of patience from leadership, and a different kind of honesty from the team running them. Early in my career, I built a website from scratch because the budget for a proper one did not exist. That experience taught me something that has stayed useful for twenty years: constraints force you to think differently, and thinking differently is where the real gains come from. The teams I have seen do experimentation well are the ones that treat it with the same rigour they apply to their core channels, defined hypotheses, clear success criteria, agreed timelines, and a genuine willingness to kill things that are not working.

The ones that do it badly treat experimentation as a skunkworks activity, underfunded, under-governed, and quietly abandoned when the quarter gets tight.

How to Structure the Balance Without Losing Either

The most practical framework I have seen work is a deliberate budget and resource split, formalised and protected at the planning stage rather than negotiated quarter by quarter. The specific ratio matters less than the fact that it exists and is defended. Some teams work on a 70/30 model: 70% of marketing investment goes into proven, optimised channels, and 30% is ring-fenced for experiments. Others run 80/20. The number is less important than the governance around it.

What kills the balance is treating the experiment budget as a reserve that can be raided when the core channels underperform. That is the most common failure mode I have seen. The moment the pipeline number looks shaky, the experiment budget gets pulled into paid search to make up the shortfall, and the experiment dies before it has produced any signal. Six months later, the team concludes that experimentation does not work for their business, when what actually happened is that they never gave it a real chance.

Protecting that budget requires a specific kind of leadership cover. The CMO or VP of Marketing needs to be willing to hold the line when the CFO asks why 20% of the marketing budget is not contributing to this quarter’s pipeline. That conversation is easier to have when the experiment portfolio is documented clearly, with hypotheses, timelines, and expected learning outcomes, rather than a vague collection of things the team wanted to try.

Forrester has tracked how global and regional marketing operations teams design their decision-making structures, and one consistent finding is that the teams with the clearest governance models are the ones that sustain both optimisation and strategic investment over time. Governance sounds bureaucratic, but in practice it just means having a written answer to the question: who decides what gets tested, how long it runs, and what counts as success?

The Problem With Experiments That Are Too Small to Matter

One of the more frustrating patterns in mature SaaS marketing is the underpowered experiment. The team agrees to run an experiment, allocates a small slice of budget, runs it for six weeks, and then looks at the results. The results are inconclusive. The team concludes the idea did not work and moves on.

What actually happened is that the experiment was never designed to produce a clear answer. The budget was too small to reach statistical significance. The time window was too short to account for natural variation in the market. The success criteria were defined loosely enough that the team could interpret the results either way depending on what they wanted to believe.

This is not experimentation. It is the theatre of experimentation, and it is arguably worse than not running experiments at all, because it produces false confidence that you have tested something when you have not.

I spent time judging the Effie Awards, which evaluates marketing effectiveness at a high level. One of the things that becomes clear when you read hundreds of submissions is how rarely teams can articulate what they actually tested, what they expected to find, and what changed as a result of what they found. Most “experiments” in marketing are post-hoc rationalisations of things that happened, not prospective tests of defined hypotheses. The discipline of writing down what you expect to happen, before you run the test, is simple and almost nobody does it consistently.

If your team is going to run experiments that are worth running, they need to be sized appropriately. That means being honest about what budget and time commitment is required to produce a meaningful result, and then either committing to that or not running the experiment at all.

When the Ratio Should Shift

The 70/30 or 80/20 split is not a fixed rule. There are conditions under which you should be running more experimentation, and conditions under which you should be doubling down on optimization.

You should shift toward more experimentation when your core channels are showing signs of saturation. If your CPL in paid search has been creeping up for three consecutive quarters, if your email engagement rates are declining despite list hygiene improvements, or if your win rates in the segments you have historically owned are starting to soften, those are signals that the market is changing around you. Optimising harder in that environment produces diminishing returns. What you need is a new motion, and that requires investment in experimentation before the existing channels deteriorate further.

You should shift toward more optimization when you have a clear growth lever that is underperforming relative to its potential. If your product has strong retention metrics and your NPS is high, but your paid acquisition is inefficient, the highest-value activity is probably tightening the acquisition engine rather than exploring new channels. How you allocate budget across the marketing mix is one of the most consequential decisions in a scaling SaaS business, and it should be revisited with fresh eyes at least twice a year.

The mistake is treating the ratio as permanent. I have seen teams that ran a 70/30 split for three years without reviewing whether it still made sense for where the business was. By year three, the market had shifted, the competitive set had changed, and the core channels that justified the 70% allocation were no longer as efficient as they once were. The team was optimising a declining asset while the experiment budget sat underutilised because nobody had updated the hypotheses.

Building a Team That Can Do Both

Optimization and experimentation require different skills, and to some extent, different personalities. The person who is excellent at tightening a paid search account, finding marginal gains through bid adjustments and negative keyword management, is not always the same person who is energised by the ambiguity of a new channel test. Both are valuable. Treating them as interchangeable is a management mistake.

In a mature SaaS marketing team, the most effective structure I have seen is one where there is a clear owner for the core channel performance and a separate owner for the experiment portfolio. They do not need to be different people if the team is small, but the responsibilities need to be explicitly separated. Otherwise, the person responsible for hitting the quarterly pipeline number will always deprioritise the experiment when the two come into conflict, which is rational from their perspective but bad for the team’s long-term trajectory.

There is also a cultural dimension to this. Teams that do experimentation well have normalised the idea that most experiments fail, and that a failed experiment that produces clear learning is more valuable than an inconclusive one. That sounds obvious, but it requires active reinforcement from leadership. If the only experiments that get celebrated are the ones that work, the team will unconsciously start running safer tests with more predictable outcomes, which defeats the purpose.

Early in my career at lastminute.com, I ran a paid search campaign for a music festival that generated six figures of revenue within roughly a day. It was a relatively simple campaign, but it worked because the conditions were right and we moved quickly. That kind of result is memorable, but it is not the norm. Most of the campaigns I ran that taught me the most were the ones that did not work as expected, and the learning came from being honest about why.

The broader context for how marketing operations teams build these capabilities is worth exploring in more depth. If you are working through how your team is structured around performance and innovation, the Marketing Operations hub covers the organisational and process questions that sit underneath the strategic ones.

Measurement Frameworks That Work for Both

One of the reasons experimentation gets deprioritised is that most SaaS marketing measurement frameworks are built entirely around optimisation metrics. Pipeline contribution, CAC, MQL volume, conversion rates at each stage of the funnel. These are the right metrics for a mature, optimised channel. They are the wrong metrics for an experiment that is trying to establish whether a new motion is viable.

Experiments need their own success criteria, defined before the experiment runs. Those criteria might include leading indicators rather than lagging ones. For a content-led experiment targeting a new segment, the right early metric might be engagement depth or return visit rate, not pipeline contribution. For a community experiment, it might be membership growth and participation rates. For a partnership motion, it might be co-marketing activity volume and referral traffic quality.

The danger is using the same measurement framework for experiments as you use for core channels and then concluding that experiments do not perform. Of course they do not, by those metrics, in the early stages. That is not a finding. That is a measurement error.

What works is a two-tier reporting structure: core channel metrics reported on the standard cadence, and experiment metrics reported separately with explicit reference to the original hypothesis and timeline. This keeps the two activities visible to leadership without forcing a direct comparison that the experiment will always lose in the short term.

Forrester has written about how marketing operations functions have evolved to handle exactly this kind of measurement complexity, and the consistent theme is that the teams that get it right are the ones that build separate reporting structures rather than trying to force everything into a single attribution model. Attribution models are useful, but they are a perspective on reality, not reality itself. Treating them as the final word on what is working is one of the more common and costly mistakes in SaaS marketing.

If you are building or refining the operational infrastructure that sits behind these decisions, it is worth spending time with the broader thinking on marketing operations as a discipline. The measurement and governance questions around experimentation do not exist in isolation from how the team is structured, how data flows through the organisation, and how decisions get made.

The Honest Conversation Most Teams Are Not Having

There is a version of this discussion that stays at the level of frameworks and ratios and governance models, and that is useful as far as it goes. But the more honest conversation is about what happens when the growth rate slows and the pressure to perform in the short term intensifies.

When a SaaS business is growing fast, there is tolerance for experimentation because the core numbers are good. When growth slows, the instinct is to cut the experiment budget and put everything into the channels that are working. That instinct is understandable, but it is often exactly wrong. Slow growth in a mature SaaS business is frequently a signal that the existing channels are saturating, not that they need more investment. Cutting experimentation at that point is cutting the only activity that has a chance of finding the next growth lever.

The teams that handle this well are the ones that have built the case for experimentation in advance, before the growth slowdown, when there is still room to have a strategic conversation rather than a crisis response. That means documenting the experiment portfolio clearly, communicating the expected timeline for results, and making sure leadership understands that the value of experimentation is not visible in the next quarter’s numbers.

It also means being genuinely honest when experiments are not working. The temptation to keep a failing experiment alive because it represents a bet the team is emotionally invested in is real, and it is costly. A clean kill decision on a failing experiment, with documented learning, is more valuable than a zombie experiment that consumes resources without producing signal.

Mature SaaS marketing is not about choosing between experimentation and optimization. It is about building a team and a system that can do both deliberately, with appropriate resources, clear governance, and honest measurement. That is harder than it sounds, but it is the work.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions