What We Found Reading 600 SaaS Pricing Pages

Over the past month, our pricing extraction pipeline parsed 600 competitor pricing pages across 11 SaaS categories — CRM, marketing automation, analytics, dev tools, design, HR tech, fintech infrastructure, security, vertical SaaS, productivity, and a long tail of niche horizontal tools. The data was a byproduct of running competitive analyses for our customers, not a curated study, which means it reflects what real companies are actually shipping in 2026, not what pricing thought-leaders write about.

Six patterns repeated across categories. Some confirm what every founder already suspects. A couple are surprising and have implications for how new SaaS companies should think about packaging.

This is what we found.

The data, briefly

The 600 pages came from competitive analyses customers ran between mid-March and mid-April 2026. The companies under analysis ranged from Series A startups to public-company carve-outs. Every page was fetched, normalized, and passed through our pricing extraction pipeline — a three-stage cascade that first tries structured HTTP parsing, then falls back to a headless-browser render for JavaScript-heavy pages, then uses an LLM "hunter" to recover data from the hardest cases. We dropped any page where extraction confidence was low after all three stages (about 8% of attempts), leaving a clean dataset where we could compare apples to apples.

A few caveats worth surfacing upfront: the sample is biased toward English-language sites, B2B (about 90%), and toward companies our customers consider competitively relevant — meaning the truly long-tail SaaS market is underrepresented. Despite that, the patterns held across categories with surprising consistency.

Pattern 1: Enterprise pricing is hidden 64% of the time

The headline number that surprised us: 64% of SaaS companies in the sample do not publish their enterprise tier pricing. They publish prices for the entry-level and mid-tier plans, then label the top tier "Custom," "Enterprise," "Contact Sales," or "Let's talk."

This is up from what felt like 50–55% three years ago, based on our memory of the same exercise pre-pipeline. The shift is real and it has a clear cause: enterprise buyers do not bounce on price-hidden pages, and revealing enterprise pricing turns into a procurement disadvantage. If your enterprise plan is $48,000/year and the next-cheapest competitor is $39,000/year, the cheapest competitor wins on the first round of vendor short-listing — even when the value gap goes the other way.

The exceptions are revealing. The 36% of companies that do publish enterprise pricing fall into three groups:

• Self-serve-first companies with a strong free or low-cost tier that funnels customers into upgrade paths organically (Notion, Vercel, Linear). Their enterprise tier exists but most ARR comes from below it.
• Category leaders confident enough that price transparency works as a moat (HubSpot publishes detailed enterprise pricing because it makes their customer-acquisition motion harder to undercut).
• Niche vertical SaaS where the buyer pool is small enough that price discovery happens through industry word-of-mouth anyway, so hiding the number changes nothing.

For everyone else — the broad middle — hiding enterprise pricing is now the default. If your strategy is "we will publish ours because transparency is good," you are deviating from the category norm. That is sometimes the right call, but it is no longer the safe one.

Pattern 2: Freemium is shrinking, free trials are growing

A second clear shift: the share of pages offering a true free tier (zero-cost permanent plan) dropped to 31% of the sample, while companies offering a 14- or 30-day free trial without a permanent free tier grew to 47%. The remaining 22% offer neither — paid only, with a money-back guarantee or a sales-led demo as the friction-reduction mechanism.

Three years ago, freemium felt like the default in B2B SaaS. The shift away from it tracks with a broader cost-discipline narrative we have heard repeatedly from customers: freemium tier abuse, support cost, and infrastructure cost have all gotten worse, and the conversion rates from free to paid have not improved. AI-related cost especially — companies offering even modest AI features in their free tier discovered that 5% of free users consume 80% of the inference budget.

The companies still running freemium tend to be those where:

1. Network effects make free users valuable to paid users (Slack, Figma).
2. Marginal cost is genuinely near zero (Notion, Linear before they shipped AI features).
3. The free tier serves a content/distribution role more than a product-trial role (Loom, Calendly).

If your product does not fit one of those three conditions, the data suggests freemium is no longer worth the support and infrastructure cost. A 14-day trial with a credit card collected at signup is becoming the new default — particularly for companies under $10M ARR where every dollar of acquisition cost matters.

Pattern 3: "Talk to sales" buttons are doubling, and they are a signal

Closely related to Pattern 1: the number of "Contact Sales" / "Talk to sales" / "Get a Demo" CTAs on pricing pages roughly doubled in the past three years. Many pricing pages now have no self-service buttons at all on the highest tier. The path to purchase the top plan is a sales conversation, full stop.

This matters as a competitive signal in two ways.

For buyers: a pricing page that sends you to sales for everything above $99/month is telling you the company is going through a deliberate trade — they are sacrificing high-velocity self-service revenue for higher ACV per deal. That is a strategic choice, not a product limitation. It usually means the product is more enterprise-capable than the marketing suggests, but also that the buying process will be longer and that price negotiation is expected.

For sellers: matching the category norm here protects ACV. The companies in our sample that resisted this trend and kept self-service buttons on their highest tier showed lower median deal sizes (estimated through public revenue disclosures and headcount) — sometimes 30–40% lower than their nearest competitor at similar revenue. Self-service pricing transparency is a feature when your buyers want it; it is a price ceiling when they would have paid more.

Pattern 4: Per-seat pricing still dominates, but cracks are visible

Per-seat pricing remained the most common monetization model in the sample at 58% of pages. This is high, but it is down from roughly 70% three years ago. The companies moving away from per-seat fall into two camps:

Usage-based pricing has grown to about 22% of the sample, almost entirely concentrated in dev tools, infrastructure, and AI/ML categories. Vercel, Modal, Replicate, OpenAI's API tier — all priced on usage. The pattern is clearest in categories where seats are not a meaningful unit of value (an API does not have "users" the way a CRM does).

Hybrid pricing — a small base fee plus usage component — accounts for the remaining shift, around 15%. This is the format we expect to grow fastest in 2026, particularly for AI-feature-heavy products. The pure-usage model is hard to forecast against; the pure-seat model under-monetizes power users. Hybrid splits the risk.

The cracks in per-seat are not yet a collapse. Per-seat is sticky because it is easy to explain, easy to budget against, and easy to grow with team size. But every CFO we hear from is increasingly suspicious of per-seat pricing for tools where utilization is uneven — exactly the case for most AI tooling, where 10% of seats produce 90% of the value.

If your product fits the "AI tool with uneven utilization" pattern, per-seat is going to keep losing ground in your category over the next 24 months. The companies experimenting with hybrid models now will set the new norm.

Pattern 5: Annual discounts have converged on 17–20%

In categories where companies offer both monthly and annual billing, the annual discount has converged on 17–20% off for 80% of the sample. Three years ago this was much wider — discounts ranged from 10% (a token gesture) to 35% (an aggressive prepay incentive). The convergence suggests companies have benchmarked each other into a narrow band.

Outliers exist on both ends:

• Below 15%: companies confident their churn is low enough that annual prepay does not need to be incentivized heavily. Often category leaders.
• Above 25%: companies running aggressive cash-flow strategies, often venture-backed, often in growth-at-all-costs mode despite the broader discipline shift.

For founders setting an annual discount, the data is clear: 17% is the floor that signals you are taking the discount seriously. 20% is the comfortable middle. Above 25% you are signaling cash flow stress to sophisticated buyers. Below 15% you are signaling that you do not really want annual prepay, which is a strange thing to signal.

Pattern 6: Add-ons are proliferating, and they are mostly AI

The most subtle pattern in the data: the number of separately priced add-ons per pricing page has grown sharply. Three years ago, most SaaS pricing pages had 0–1 add-ons (typically a "premium support" SKU). Today, the median is 2 and the 90th percentile is 5+ add-ons.

Almost all the new add-ons are AI features. The pattern across categories is consistent:

• A core plan that includes the product's traditional functionality
• An "AI Insights" or "AI Co-pilot" add-on at $X/month/seat or $Y per 1,000 actions
• An "Advanced Security" add-on (often SOC 2 reports + SAML + audit logs) at $Z/month
• A "Premium Support" add-on
• Increasingly, an "API access" add-on if the base plan does not include API

This pattern reflects a tension companies are working through: AI features are expensive to deliver, customers are willing to pay for them but not always, and bundling them into base plans risks margin destruction. Separating them as add-ons solves the problem at the cost of a more confusing pricing page.

It also creates a competitive signal. A company adding an AI add-on for $19/seat is signaling their inference cost per active user is roughly in that range — useful information when modeling your own pricing, particularly if you are in the same category. We have seen this in our customer data: pricing teams use competitor add-on prices as anchors for their own AI-feature monetization decisions, with surprisingly little discussion of whether the anchor is correct.

What this tells us about the next 24 months

Pulling the patterns together, the picture of where B2B SaaS pricing is heading:

1. Pricing pages are becoming less transparent on average, especially at the top end, as enterprise buyers price-shop and as procurement-led buying grows back to pre-2022 levels.
2. Freemium is in decline, replaced by trial-based acquisition with credit card collected upfront.
3. Sales-led motion is recovering ground lost to product-led growth during 2020–2022, particularly above $50k ACV.
4. Per-seat pricing will keep losing share to hybrid models, especially in AI-heavy categories, but it will not collapse.
5. Annual discounts have stabilized at 17–20% — this is a solved problem and not where pricing innovation will happen.
6. Add-on proliferation will continue as a way to monetize AI features without destroying base-plan margins, until the category settles on a new bundling norm.

The companies that get this right are paying close attention to what their direct competitors actually publish — not what pricing playbooks recommend in the abstract. The 17–20% annual discount band is real because companies looked at each other and converged. The 64% enterprise-hidden rate is real because companies looked at each other and converged. Pricing decisions in B2B SaaS are increasingly downstream of competitive observation.

That is also why we built Seeto the way we did. Pricing intelligence is one of the four dimensions our analysis covers — the others being features, SEO, and messaging — but it is the one customers tell us they look at most often. Knowing how your competitors package and price, in structured form rather than as 12 open browser tabs, is most of the work of designing your own pricing.

If you want a structured pricing snapshot of your top 5 competitors in the next 5 minutes, run a free analysis. We will not match the depth of a dedicated pricing consultant — but we will give you a current-state picture you can act on, which most of the consulting deliverables we have read do not.

---

Pattern data is drawn from competitor pricing pages parsed by Seeto's analysis pipeline between March 15 and April 15, 2026. All percentages are approximate and reflect the specific composition of our customer-driven sample, not a census of the SaaS market.