What 852 pricing pages taught us about how software is sold

We have now read, structured, and fact-checked the pricing page of every product in the SaaSTracker database: 852 tools across 49 categories, from CRMs to call tracking to notification infrastructure. A pricing page is the one document where a vendor's actual strategy is written down in public, and reading hundreds of them side by side reveals patterns no single page shows. This piece is the findings.

The headline numbers first, all computed across the tools we track. 466 of the 852 have a genuine free plan. 809 offer a free trial. Only 26 offer neither, which means trying software before paying for it is now close to universal. Among the 523 tools publishing a flat dollar entry price, the median advertised entry point is $38 a month; a quarter start under $19 and a quarter above $89. 220 products meter something with credits. 221 price per seat. 33 publish no usable entry price at all, and 115 carry expiry or rollover language somewhere in the fine print.

Those are the aggregates. The interesting part is what they mean.

The free plan follows the marginal cost, not the generosity of the vendor

Vendors talk about free tiers as philosophy. Product-led growth, community, giving back. The data says something plainer: whether a category offers free plans is almost entirely predicted by what a free user costs the vendor to serve.

Look at the extremes across our 49 categories:

Category	Products with a free plan
Email verification	17 of 17
Subscription billing	17 of 17
B2B data	24 of 26
Email service providers	20 of 22
Scheduling	20 of 22
LinkedIn outreach	3 of 19
Visitor identification	3 of 18
Local SEO	1 of 12
Call tracking	1 of 13
Sales calling	1 of 15

Every single email verification tool we track has a free plan, because verifying an email is a cheap API call and a thousand free checks cost the vendor almost nothing. Every subscription billing tool has one, because billing platforms take a percentage of the money they move; a free customer processing zero dollars costs them nothing today and becomes revenue automatically the day the customer earns any. Free entry is not generosity in either category. It is a rounding error with a funnel attached.

Now the bottom of the table. One sales calling tool in fifteen offers a free plan, because every call a free user makes consumes telephony minutes the vendor pays a carrier for. Call tracking is one in thirteen for the same reason: the phone numbers themselves rent by the month. LinkedIn outreach manages three free plans in nineteen products, and there the cost is not minutes but risk, since every automated account is a liability the vendor has to manage whether or not it pays.

This explains free tiers better than any vendor's stated philosophy, and it gives you a useful prior. When a category is structurally cheap to serve, expect a free plan and be suspicious of any vendor who withholds one. When it is structurally expensive, a generous free tier is either venture capital burning or a limit you have not found yet.

The modern free plan is an appointment, not a product

Having read several hundred free-plan descriptions in a row, we will state the house position bluntly: the contemporary free tier is calibrated to cost you an afternoon, not to run your business. It is convincing for exactly as long as it takes you to build something inside it, and the critical function, the export, the automation, the report, the second workflow, sits one tier up, waiting for the moment you are committed.

The gates are easy to spot once you know the shape. Typeform's free plan allows unlimited forms and, since February 2026, ten responses a month across the whole account, with no logic and no integrations. Its own limit behavior tells you what the plan is for: hit the cap and collection pauses, and visitors' answers are simply not recorded. SurveyMonkey's Basic plan goes a step further and restricts how many of your collected responses you can even view, so data gathered past the line is held until you upgrade. Your respondents answered; the answers are behind the paywall.

The export gate is subtler and meaner. Forminit's free plan takes your form submissions happily, but CSV export is gated to the $19 Pro plan, so the free tier can collect your data and cannot give it back in bulk. And the automation gate is the classic of the genre: Kit's free Newsletter plan is genuinely huge on capacity, up to 10,000 subscribers with unlimited broadcasts, but it includes exactly one basic visual automation. You can have the audience for free. The moment you want the machine to work while you sleep, which is the entire promise of the product, you are on a paid plan. Mailchimp's free tier plays the same note in a different key: 250 contacts, no send scheduling, no multi-step automation, a plan trimmed twice since its generous years.

Fairness requires the counterexamples, because genuinely usable free plans exist and are worth seeking out. Tally gives away unlimited forms and unlimited submissions, with conditional logic, calculations, file uploads, and Stripe payments all on the free plan; the paid tiers sell branding removal and governance, not your own data back. Fillout's free plan takes 1,000 responses a month with unlimited seats and payment collection included. The pattern among the honest ones is consistent: they charge for polish and scale, not for the verb you came to do.

The test, then, takes one minute. Find the thing you actually need the tool for, and check whether the free plan can start it, finish it, and export it. If any of those three is missing, you are not evaluating a product. You are attending an appointment the vendor booked for you.

Credits are where pricing goes to hide

220 of the 852 products we track meter something with credits, and the range of what a credit buys is now absurd: an email verification, a phone number reveal, an AI agent session, an SMS segment, a workflow run, a PDF export. Credits are not inherently dishonest. They are a reasonable way to price heterogeneous usage. But our position after reading all 220 is this: a credit price you cannot convert into a unit cost in under a minute is a design choice, not an accident, and it is the single fastest legitimate reason to close a pricing tab.

The honest version exists. Textmagic sells prepaid texting credit with no subscription at all: roughly $0.049 per US SMS on its own routes, about $0.01 if you connect your own carrier, and the credit never expires. No monthly forfeiture, no rollover cap, nothing to cancel. You can compute your cost per message from the pricing page in seconds, which is the entire point.

The opaque version also exists. Textdrip advertises the lowest headline credit rate in its category, around $0.012 per credit, and then applies a multiplier most buyers miss: on the $19.99 Spark Starter plan an SMS costs two credits, while on the $34.99 Growth Gear plan the same SMS costs one. The cheapest plan is literally double the per-message price of the middle plan, and once registration fees and the add-ons Growth Gear bundles are counted, the entry tier is more expensive than the tier above it for almost any sustained use. Nothing on the plan card says so. The multiplier lives in the fine print, doing exactly the work it was designed to do.

Then there is expiry. 115 of the 852 products carry expiry or rollover language in their pricing fine print, and the terms range from civilized to confiscatory: SimpleTexting rolls unused credits into the next month, SlickText publishes a full year of rollover on annual billing, and at the other end sit vendors whose paid credits evaporate at the end of each billing month. We have written the long version of that investigation, with named terms across categories, in our piece on the fine print of SMB SaaS, so here we will only restate the conclusion: paid credits that expire monthly are the worst standing pattern in small-business software. You paid for capacity, the vendor kept the money, and the meter quietly reset. Before buying any credit bundle, ask two questions in order. What is one credit in units I understand, and what happens to the ones I do not use? A vendor who makes either answer hard to find has answered a third question you did not ask.

The advertised price is a floor stated as a promise

Across the 523 tools publishing a flat entry price, the median is $38 a month, with quartiles at $19 and $89. Those numbers are real; the vendor will genuinely charge them, for the configuration on the banner. The recurring problem is that the banner configuration and a working configuration are different products, and the distance between them is bridged by add-ons.

The pattern we see most often is a core plan that is honestly priced while the required extras double the invoice. Pipedrive is the cleanest case study: the Lite tier is $14 per user per month on annual billing, a fair price for a good pipeline tool, but chat capture, web visitor tracking, email marketing, and document features are all separate add-on modules billed on top. By the vendor's own published add-on prices, a five-seat team that wants email sync, automation, chat capture, and email marketing lands closer to $250 a month than the $70 the entry price implies. In texting, Salesmsg advertises $25 a month, which buys exactly one seat and one phone number; a five-person team adds $40 a month in seats before anyone sends a message, plus $5 per additional number, plus carrier fees. In marketing automation, EngageBay's Pro tier permits 50,000 contacts but only 50,000 branded emails a month, which is one send to your full list; the workaround everyone uses is connecting an external sending provider, a second bill the pricing page never mentions.

None of this is fraud, and most of it is disclosed somewhere. But an advertised price that requires three additional line items before the product does its job is not a price. It is a floor stated as a promise. We took apart the entry-tier version of this problem, where the first usable plan sits one tier above the advertised one, in our analysis of what "starting at" really costs; the add-on version is the same trick performed sideways, and the defense is identical. Price the configuration you will actually run, not the one on the banner.

Where per-seat survives and where it is dying

221 of our 852 products price per seat, and the line between where per-seat holds and where it is retreating is one of the clearest in the whole dataset: the seat survives where it maps to a human doing the work, and gives way to usage metering wherever the work is done by the tool.

CRM is the seat's home territory, and it is not leaving. Close, Pipedrive, and Less Annoying CRM all price per user, and it makes sense, because a CRM seat is a rep, and a rep is a unit of selling capacity. The same logic holds in per-user business texting like Heymarket, where a seat is an agent answering a shared inbox. When the human is the engine, charging per human is coherent.

But watch what happens inside those same products the moment the tool starts doing the work. Close is per-seat for the humans and usage-billed for the machine: calling minutes, SMS, enrichment, and AI credits all meter separately, and a rep doing three hours of talk time a day adds roughly $70 to $80 a month on top of the seat. Attio charges $29 per user and then runs a separate workspace-credit meter for AI and enrichment usage, with extra credit packs sold monthly; its own pricing notes make clear the packs are a real line item, not an edge case. Freshsales includes Freddy AI sessions once, then charges $49 per hundred beyond the allowance. Nutshell allots AI outcomes per tier. The seat is becoming the membership fee, and the meter is becoming the bill.

Outside the CRM, in categories where no human sits in the loop at all, the seat is simply gone. Email verification prices per check. Sending infrastructure prices per email; Brevo built its entire pitch on metering emails sent rather than contacts stored. Enrichment prices per record, notification platforms per message or per workflow run. Nobody charges per seat for an API, because a seat that never touches the product is an obviously silly unit, and buyers stopped accepting it. The direction of travel is one way. Work is migrating from reps to software, and pricing is following the work. If you are signing a per-seat contract today, the question to ask is not what a seat costs now but what the meters bolted to its side will cost when the AI features you were sold actually get used.

What a founder should do with all of this

Compressed to a single paragraph, the five-minute pricing-page read goes like this. Find the billing toggle and write down the monthly price, not just the annual one, because the monthly price is the price of leaving. Find the unit, the seat minimums, and the multipliers, and multiply by your real team. Read the entry tier's limits column, not its feature list, against the three tasks you are actually buying the tool for; if one of them lives a column to the right, that column's price is your price. Convert any credit to a unit cost and get the expiry terms in writing. And add every mandatory extra, onboarding, add-on modules, pass-through fees, into a year-one number before you compare anything against anything. Five minutes per tool. It is the highest-yield diligence available to a small company, and after 852 pricing pages we can report that almost every unpleasant surprise in software billing was printed on the page all along, one click below the banner.

The short version

We read all 852 pricing pages so you can skim them smarter. Free plans track the vendor's marginal cost, not their values: universal where serving you is free, nearly absent where it is not. Most free tiers are appointments engineered to gate the export, the automation, or the report at the moment of commitment, though honest ones like Tally's exist. Credits are the fog of the industry, at their worst when a multiplier hides the unit cost or an expiry clock confiscates what you prepaid. The $38 median entry price is a floor that add-ons routinely double, and per-seat pricing is retreating to the places where a seat still means a person. Read the limits, price the meters, and budget the column to the right.