Industry
B2B SaaS marketing
The conversion you choose decides what the algorithm optimises toward, and most accounts choose wrong.
B2B SaaS has the sharpest conversion-definition problem in paid media. Free trial and demo request are both valid conversions that select for completely different buyers, and revenue arrives months after the click, outside every attribution window. Getting those two decisions right matters more than anything else in the account.
What differs
What makes B2B SaaS paid acquisition different?
Revenue arrives months after the click and the conversion you optimise toward selects the buyer, so measurement design decides performance more than bidding does.
The defining problem is that you must choose what to call a conversion before the algorithm can optimise, and in B2B SaaS every available choice is wrong in a different way. A free trial signup is plentiful, fast, and heavily weighted toward individuals evaluating for personal use or curiosity. A demo request is scarcer, slower, and weighted toward buyers with budget. Optimise toward the first and you get volume that never buys; optimise toward the second and you may starve the bidding of data entirely.
The second defining problem is time. Revenue arrives weeks or months after the click, far outside any conversion window, so the platform will never see the outcome unless you send it back. Without offline conversion import the account is optimising toward a proxy chosen for convenience, and the gap between that proxy and revenue is where most wasted B2B spend lives.
Third, the buying decision is made by several people who do not all appear in your analytics. A user researches, a manager evaluates, a finance approver signs. Attribution models assume a person, not a committee, so the model will always describe less than what happened.
Finally, volume is thin. Most B2B SaaS accounts produce few enough conversions per campaign that automated bidding struggles to learn, which changes what structures are viable. Consolidation matters more here than segmentation, which is the opposite of the instinct most accounts arrive with.
Intent
What are the intent patterns in B2B SaaS search?
Category, alternative and competitor queries dominate, with a large share of the research done anonymously long before anyone is willing to identify themselves.
Category and problem queries
People searching for a class of software rather than a brand. The main acquisition volume, contested by review sites and directories that frequently outrank every vendor.
Competitor and alternative queries
Searches for a rival by name, or for alternatives to one. Very high intent, cheaper than category terms, and usually the best value in the account provided the landing page is an honest comparison rather than an attack.
Integration and compatibility queries
Whether the product works with a specific tool the buyer already runs. Overlooked, high intent, and often a hard requirement rather than a preference.
Pricing queries
Searched heavily and frequently unanswered by vendors, which pushes buyers to review sites that guess. A pricing page that explains structure captures people who are otherwise still comparing.
Jobs-to-be-done and workflow queries
How to do a specific task, where the software is one possible answer. Top of funnel, valuable for content, and rarely worth a direct demo ask.
Negative keyword themes this vertical needs
Themes that matter here specifically. A generic negative list will not contain them, and each one is a recurring source of spend that looks like relevant traffic.
- Free, open source and forever-free
- Searchers explicitly filtering for free tools rarely become paying accounts, and the terms sit directly alongside category queries.
- Careers and hiring
- People searching to work at a software company, or for roles using a tool, share vocabulary with buyers and convert to nothing.
- Support, login and documentation
- Existing users looking for help or a login page are already customers. Paying to serve them ads is a common and invisible leak.
- Tutorials, courses and certification
- Learning-intent queries around a category are enormous in volume and almost never purchase intent.
- Consumer and personal-use variants
- Many B2B categories have a consumer analogue using identical words, and consumer intent will not convert on a business plan.
Measurement
What counts as a conversion in B2B SaaS?
Qualified pipeline, not trials or demos. Both upstream signals are easy to buy and select for very different and unequally valuable populations.
Trial and demo select different people
A trial selects for hands-on individual users; a demo selects for buyers with budget and a timeline. They are not two routes to the same person, and choosing one as the optimisation target chooses which population the algorithm hunts.
The real conversion is qualified pipeline
Neither trials nor demos are revenue. Qualified pipeline is the first signal that correlates with it, and it lives in the CRM rather than in the ad platform, so it only reaches bidding if you send it.
The sales cycle outlives every attribution window
Deals closing months after the click will never appear in a default window. The campaigns that started the relationship therefore look worthless and the ones that closed it look brilliant, which is how good top-of-funnel activity gets cut.
Buying committees break person-based attribution
Several people research, and only one converts. Every attribution model will credit a fraction of what actually happened, and account-level rather than person-level thinking is the only honest response.
Thin volume starves the bidding
Automated bidding needs conversions to learn from, and many B2B accounts do not produce enough per campaign. That argues for consolidation and for optimising toward a higher-volume upstream event with the real outcome imported behind it.
Constraints
What constraints apply to B2B SaaS advertising?
Thin conversion volume starves automated bidding, buying committees are invisible to attribution, and no platform can target a company decision.
No platform can target a buying decision
You can target job titles on some platforms and infer intent on others, but nobody can target a committee reaching a decision. Every targeting option is a proxy for the thing you actually want, and treating proxies as the thing is how B2B budgets get spent confidently on the wrong audience.
Review sites and directories own much of the category SERP
Software review platforms frequently outrank every vendor on category terms, and they monetise by selling the same buyer to several of them. You are bidding into an auction with participants whose economics differ fundamentally from yours.
Competitor bidding has trademark limits
Bidding on a competitor's name as a keyword is generally permitted, while using their trademark in ad text usually is not, and platforms operate complaint processes that can result in disapproval. Build the comparison honestly on the landing page rather than in the headline.
Source: Google Ads trademark policy
Thin data limits what structures are viable
Segmentation that would be obviously correct at higher volume splits an already sparse conversion count into campaigns that never exit learning. The constraint is real and it argues against the structure most accounts instinctively want.
What good looks like
What does a good B2B SaaS account look like?
Optimised toward qualified pipeline imported from the CRM, structured around category and competitor intent, with realistic expectations of thin data.
- Qualified pipeline imported from the CRM as the outcome signal, with trials or demos used as the upstream optimisation event when volume requires it.
- One deliberate choice between trial-led and demo-led acquisition, made on which buyer you actually want rather than on which number looks better.
- Competitor and alternative campaigns run separately, landing on honest comparisons rather than attack pages.
- Consolidated campaign structure, because thin conversion volume punishes segmentation that would be correct at scale.
- Pricing structure explained on the site, so review sites are not the ones answering the question for you.
- Account-level rather than person-level reporting where the CRM supports it, since the buying committee never fits a person-based model.
- Realistic reporting cadence, because a sales cycle measured in months cannot be judged weekly without reading noise as signal.
When this vertical does not need us
If you have no CRM connection and no intention of building one, we are the wrong choice, because we would be optimising toward form fills and calling it performance. If your product has genuine product-led growth and self-serve conversion at volume, in-house growth engineering will beat any agency on iteration speed and you should hire for that instead. And if you are pre-product-market-fit, paid acquisition will buy you expensive confirmation of an unclear proposition rather than growth.
Related services:performance marketing servicesmarketing analytics services
Questions
What else comes up in B2B SaaS marketing?
Whether to run trials or demos, how to bid on competitor names, dealing with low conversion volume, and measuring long sales cycles.
Should we run free trials or demo requests?
Whichever matches the buyer you want, decided deliberately rather than by default. Trials suit products a single user can evaluate and adopt; demos suit considered purchases with budget holders and procurement. Running both without deciding which the algorithm optimises toward is the common failure, because the platform will quietly pick the cheaper one.
Is bidding on competitor names worth it?
Usually yes in this vertical, because it is high intent and cheaper than category terms. Keep their trademark out of your ad text, expect them to do the same to you, and make the landing page an honest comparison. A comparison page that concedes nothing is not believed by anyone evaluating seriously.
What do we do about low conversion volume?
Consolidate rather than segment, optimise toward a higher-volume upstream event, and import the real outcome behind it so the platform eventually learns from what mattered. Fighting the constraint with more granular campaigns is the instinct and it makes the problem worse.
How do we measure a six-month sales cycle?
With offline conversion import and cohort reporting rather than in-platform conversion counts. Accept that this quarter's spend is judged next quarter, set the reporting cadence to match, and resist reading weekly movement as signal when the underlying cycle is measured in months.
Should we run LinkedIn as well as Google?
If you sell to a definable job function, LinkedIn offers targeting Google cannot match, at much higher cost per click. That premium is justified when the audience is narrow and the deal is large. It is not justified when you are still learning which audience converts, because you will learn that more cheaply on search intent.
Related industries
iGaming marketing
The account structure is dictated by policy before a single campaign is built.
Running in this vertical?
Tell us what the account looks like now. We will tell you which of these constraints is actually costing you money, and whether the answer is us or a specialist.