Three real differences, and several imagined ones
Software marketing differs structurally from the rest of B2B in three respects, and only three.
The product can sometimes be distributed and evaluated without a human. A prospect can acquire, configure and judge the thing before speaking to anyone. No industrial component or consulting engagement works that way, and every genuinely software-specific tactic originates here.
Revenue is recurring, so the sale is never finished. A customer acquired badly does not merely fail to expand; they leave, having consumed onboarding and support cost. Marketing therefore has a legitimate stake in what happens after signature.
The product is instrumented. You can observe what a prospect actually does at a granularity no other B2B category approaches. That is a real analytical advantage, regularly wasted on dashboards nobody acts on.
What does not differ is longer than what does. Buying committees, procurement, security review, legal redlines and budget cycles are not softened by the product being software. Most wasted spend in this sector comes from treating a company as a software business when it is an enterprise sales business with software as the deliverable.
Free trials and product-led growth have preconditions
Product-led growth — using the product itself as the primary acquisition and conversion mechanism — works when a specific set of conditions holds. It is not a philosophy; it is a fit test.
- The product demonstrates value without a human, inside the trial window. Not eventually. Within days.
- A single user can get value before the team does. If the product only works with five colleagues in it, one trialist experiences an empty room.
- The user can get their own data in easily. If evaluation needs a production system connected, the trial ends at the integration step.
- The user and the buyer are the same person, or close. A developer adopting a tool on a company card is not a platform purchase requiring a security review.
- No implementation or professional services are needed to see the point.
Most B2B software fails at least one of these. When the conditions do not hold, a free trial does not become a slower product-led motion — it becomes a lead capture mechanism with a worse conversion rate than a demo request, and the prospect's first experience of your product is a blank, unconfigured version that does nothing.
The test is whether an unaided stranger can reach the moment your product is obviously useful, alone, in a week. If not, a guided pilot is the correct motion, not a retreat from best practice.
Three sales motions living on one website
Most software companies of any size run self-serve, sales-assisted and enterprise motions at once, and most websites are built as though only one exists. Self-serve signup buried behind a contact form pushes away the practitioner who would have paid immediately. A prominent trial button sends the enterprise buyer into a product experience that will not survive their security requirements.
What works is segmentation by page rather than by interruption:
- Parallel primary actions, not a hierarchy. Start free and talk to sales can sit side by side.
- Distinct destinations for distinct motions — a product page for practitioners, a solution page for a department, and an enterprise page that reads as though written by someone who has been through a procurement process.
- An enterprise evidence layer: security documentation, compliance certifications, uptime history, subprocessor lists, and a route to a real contract discussion. A buyer's security team will find this page whether or not you built it.
The fight this exposes is rarely about the website. It is about which motion the company is committed to, and the website is where an unresolved answer becomes visible to customers.
The pricing page, and the argument about it
The pricing page is the most-read page on most software websites and the most argued-about decision in most software companies. Both facts have one cause: it is the only page that forces the company to commit to a position.
What transparency buys is real. It qualifies out prospects who were never going to buy before they consume sales time, answers the question every evaluator asks first, and lets a self-serve motion exist at all. What it costs is also real. It anchors negotiations, exposes structure to competitors, and means little for negotiated enterprise agreements.
The defensible position is to publish the structure even where figures cannot be published. Say what the unit of pricing is, what drives the total up, which capabilities sit in which tier, what is charged separately, and where self-serve stops and negotiation starts. A page where every tier says contact us communicates something, and it is not confidence.
One decision belongs to marketing though it looks like product: the pricing metric. Per seat, per usage, per record, per transaction — the choice determines who inside the customer feels the cost and whether their growth grows your revenue.
Churn, and what it does to marketing's remit
In a transactional business a poorly-fitting customer is a slightly disappointing quarter. In a subscription business it is a negative-value customer: they consumed acquisition spend, onboarding effort and support capacity, then left before contributing. This is the dimension other B2B categories do not share.
Lead quality has to be measured against retention, not conversion. A channel with a strong cost per opportunity and poor twelve-month retention is destroying value while looking efficient in the acquisition report. Segmenting retention by original acquisition source is the most useful analysis available to a SaaS marketing team, and it needs customer data joined to marketing data — usually the actual blocker.
Overclaiming becomes measurable. Marketing that promises a capability the product lacks does not fail at the point of sale; it fails two quarters later as a renewal that does not happen.
Post-sale content is marketing work. Documentation, onboarding sequences and adoption guides all determine whether the revenue recurs.
None of this makes marketing accountable for gross churn, which is driven mostly by the product. It makes marketing accountable for who was brought in and what they were told.
Category creation is oversold
Creating a category — naming a new class of product and persuading the market to adopt the name — is the most glamorous strategy in software marketing and the wrong choice for the overwhelming majority of companies that attempt it.
The argument for it is coherent. Define the category and you set the evaluation criteria, and the criteria happen to describe your product.
The problems are practical. Category creation requires analysts, journalists, customers and competitors to adopt your framing — none of which you control, all of which take years and capital. More importantly, it removes your ability to capture existing demand. Nobody searches for a category that does not yet exist, so the channel where buyers with budget actually appear is closed to you by definition.
The visible successes are also a survivorship sample. The companies held up as proof were unusually well funded, often riding a platform shift that created the demand independently. The genuinely correct cases are narrow: no existing category describes what you do, buyers have no budget line, and you can fund years of education.
Where received SaaS wisdom comes from
A great deal of what circulates as SaaS best practice originates with a small number of exceptional companies in unusual conditions, and was written up after they had already succeeded.
- Cheap capital. Generous free tiers and content programmes that took three years to repay were financeable in a way they now are not.
- Empty search results. The content-moat strategy was executed when the relevant queries returned almost nothing. The same effort against today's results, with AI features occupying the top of the page, does not produce the same outcome.
- Products with inherent virality or developer distribution. Advice from them does not transfer to software one person uses alone.
Two specific pieces deserve retirement. First, that product-led growth replaces sales; most companies that grew that way added a sales organisation as soon as they moved upmarket. Second, the metric set inherited from the era — qualified leads counted as an outcome, self-reported attribution treated as measurement, a satisfaction score standing in for retention risk. Report pipeline, win rate and revenue retained by cohort, and let the rest be diagnostics.
What good looks like
The order to work in is not the order most software companies choose.
- Decide which motion you are in. Self-serve, sales-assisted, enterprise, or a defined combination with a stated boundary.
- Test the product-led preconditions honestly before building a trial. If an unaided stranger cannot reach obvious value alone in a week, build a guided pilot instead and stop apologising for it.
- Build the enterprise evidence layer once — security, compliance, data handling, uptime.
- Join marketing data to retention data. Until acquisition source can be segmented by twelve-month retention, you are optimising the wrong thing.
- Pick a few subjects you can be the definitive source on, and publish material only your company could write.
The thing to watch is the gap between how the company describes its motion and how customers actually buy. That gap opens quietly, as the company drifts upmarket while the website keeps addressing the practitioner who founded the first thousand accounts. Once deals require a security questionnaire and a procurement portal, the trial button has stopped being the answer to anything.
Frequently Asked Questions
Should we offer a free trial or a demo?
Offer a trial only if a stranger can reach obvious value alone, inside the trial window, without your data being connected or your team configured. If evaluation needs an integration, several colleagues or an implementation step, a trial gives the prospect an empty product.
Where the preconditions fail, a guided pilot converts better. Many companies should offer both — the trial for practitioners, the guided path for anything passing through procurement.
Should we publish our pricing?
Publish the structure in all cases, and publish figures wherever the product can be bought without negotiation. Structure means the unit of pricing, what drives the total up, what sits in which tier, what is extra, and where self-serve stops and negotiation starts.
Hiding pricing entirely costs you qualification and credibility, and a page where every tier says contact us reads as a warning.
Does product-led growth mean we do not need a sales team?
No, and the companies most often cited as proof built sales organisations as they moved upmarket. Product-led growth changes what sales does: the prospect arrives having already used the product, so the conversation is about expansion, security, procurement and terms rather than explaining features.
What genuinely changes is the trigger. Instead of a form fill, sales acts on usage signals. That requires product telemetry to reach the CRM.
Should marketing be responsible for churn?
Marketing should be responsible for who was acquired and what they were promised, not for gross churn, which is driven mainly by the product and the customer's own circumstances. The practical test is to segment twelve-month retention by original acquisition source: a channel producing cheap opportunities and customers who leave is destroying value while appearing efficient.
Marketing usually also owns the post-sale content that decides whether revenue recurs.