TL;DR
- OrganikPI checked 21 enterprise SaaS vendors whose own pricing pages publish no number and gate pricing behind a demo form.
- Sixteen of the 21 have a real, public price attached to them, and 11 publish it themselves, nine of those as vendor-direct AWS Marketplace listings.
- Braze's own blog post says it has no public pricing, in FAQPage structured data written for machines, while Braze lists $75,000.00 for a 12-month contract on AWS Marketplace.
- Oracle publishes $625.00 per hosted named user per month for Fusion ERP (part B91079) in a PDF on oracle.com.
- Rippling publishes at least four starting prices across regional pages in three currencies, with its UK page quoting US dollars and its Ireland page quoting British pounds, while its US pricing page shows nothing.
- Anaplan has three competing public numbers, from a $750,000 AWS listing to a median of $114,609 across 67 real purchases on Vendr, a gap of between six and seven times, and none of the three are on anaplan.com.
- This matters because AI answer engines lean heavily on third-party pages for pricing questions: VisibleIQ found 93.7% of decision-stage citations on Perplexity, Gemini and Claude pointed to third parties, and captured ChatGPT reasoning shows the model falling back to G2 when a vendor's own pricing page cannot be parsed.
In January, Braze published a blog post called “Braze Pricing: Finally, Enterprise Pricing That Makes Sense.”
It opens like this:
Budgeting for marketing technology shouldn’t feel like a guessing game. Too often, enterprise teams are forced into a long “black box” sales cycle where the final price in the contract is opaque, confusing, and unpredictable.
Braze
And then, a line later:
We believe you deserve to know exactly where your investment goes.
Braze
At the bottom of that post is an FAQ. The first question is “Does Braze have public pricing?” Here is the answer, verbatim:
No, because we don’t believe in a “one-size-fits-all” rate card. Instead, we customize your package based on your specific goals and usage.
Braze
The next question is “How much does Braze cost?” The answer runs 47 words and contains no number.
Braze also sells on AWS Marketplace.
Public. No login. Sold by Braze. $75,000.00.
So Braze published a tidy, machine-readable answer to the question “does Braze publish pricing,” and that answer is no, and it is served from a post about how you deserve to know where your money goes, and it is contradicted by Braze’s own storefront on the largest software marketplace in the world.
Nobody at Braze is lying. That’s the part worth sitting with. Every one of those artifacts was produced by someone doing their job properly.
I went looking at 21 enterprise software vendors whose pricing pages publish no number and gate everything behind a demo form. I wanted to know whether “contact sales” meant “we don’t publish a price” or something else.
It means something else. Sixteen of the twenty-one have a real, public price attached to them. Eleven publish it themselves.
The prices are on AWS
Nine of the eleven are AWS Marketplace listings where the seller is the vendor. Not a reseller, not a partner. The vendor.
| Vendor | Public price | Their own pricing page |
|---|---|---|
| Infor | $1,000,000.00 / 12mo (CloudSuite) | /pricing returns a 404 |
| Anaplan | $750,000.00 / 12mo (Connected Planning) | redirects to an office address page |
| Coupa | $150,000.00 / 36mo | talk to sales |
| Cornerstone OnDemand | $86,000.00 and $118,000.00 / 12mo | redirects to an LMS FAQ, no numbers |
| Heap | $100,000.00 / 12mo | free tier, paid tiers quote-gated |
| iCIMS | $90,000.00 / 36mo (100 employees) | branded 404, US and UK |
| Braze | $75,000.00 / 12mo | four tiers, not a number on any of them |
| Sisense | $40,000 / $69,000 / $109,000 | free trial, talk to us |
| Domo | $30,000 to $600,000 (8 tiers) | talk to sales |
I did not find these by guessing or by asking anyone. They took about a minute each.
There is a mechanical reason this happens, and it isn’t hypocrisy. To transact through AWS committed spend, you list a price. Enterprises want to buy through their existing AWS agreement because the money is already allocated. So the vendor lists. Meanwhile the website is run by marketing, whose job is to get a form filled in. The two teams answer to different people and the contradiction never surfaces internally, because nobody at the company is looking at both pages at once.
Your buyers are. B2B buyers now name generative AI among their top sources of self-guided information at every stage of the buying process, and most of the buying journey is finished before anyone fills in your form.
Oracle just publishes a rate card
Oracle is the clearest case, and the one I’d point at if you only look at one.
Oracle hosts a PDF on oracle.com called the “Oracle Fusion Cloud Service Global Price List.” It is dated May 7, 2026. It requires no login.
$625.00 per hosted named user, per month. Minimum 10 users. Part number B91079.
That is a per-user, per-month price for Oracle ERP, published by Oracle, sitting on Oracle’s own domain right now.
It took me three tries to read it. The PDF’s table layout falls apart under normal text extraction, and every price gets detached from its product name. Nothing is hidden. The columns just come out scrambled. The fix was one flag on one library. Which means anything reading that page automatically, including the crawlers feeding AI answer engines, sees a column of orphaned numbers with nothing attached to them.
The price is public. It is also, functionally, unreadable. That is a rendering and extraction problem, and it is the same category of problem as a pricing table that only exists after JavaScript runs.
Rippling publishes four prices and shows Americans none of them
Rippling’s US pricing page is quote-gated. No number.
Rippling’s other pages are not.
That is the UK site. The flag is in the header. The price is in US dollars.
That is the Ireland site. The price is in British pounds. Ireland has used the euro since 2002.
There’s an Australian page too, at A$11 per user per month, and a Canadian one at $8 plus a $40 base fee.
So Rippling publishes at least four starting prices, in three currencies, at least two of which are attached to the wrong country, and the one market that gets nothing at all is the United States.
Nobody decided this. That’s what makes it worth showing you. This is what happens when pricing lives in a CMS and nobody owns the question.
The bigger problem: there is no “the price”
Anaplan is where this stops being funny.
Ask what Anaplan costs and there are three public answers, before you even count the ones on the content farms:
| Source | Number | What it actually measures |
|---|---|---|
| Anaplan’s own AWS listing | $750,000 / 12mo | list price, one bundle |
| Anaplan’s UK G-Cloud filing | £180 to £25,000 a licence a year | list range, vendor-submitted |
| Vendr, across real purchases | median $114,609 / year (67 deals) | what buyers actually paid |
Two of those are Anaplan’s own filings, the AWS listing and the G-Cloud entry. All three are public. The gap between the AWS list price and what buyers actually pay is between six and seven times.
None of them are on anaplan.com.
These numbers don’t contradict each other. They measure different things, and every one is defensible on its own terms. That’s exactly the problem. A buyer asking a reasonable question gets a number with no way to know which of the three they’re holding, or that the other two exist.
One more wrinkle, because it matters for anyone trying to check this. The Vendr figure moves. When I first pulled it, the median was $102,032 across 51 purchases. Days later the same page said $114,609 across 67 purchases. It’s an average of closed deals, so of course it drifts. But nobody who quotes it says when they took the snapshot, which means every secondhand citation of “Vendr says Anaplan costs X” is a photograph of a moving target with no date on it. Freshness is a real ranking input for AI answers, and an undated number is the worst possible input to it.
Somebody already answered the question for you
Here’s the part I didn’t expect.
There’s a site called checkthat.ai. It has pricing pages for 5,976 brands. It’s built by an agency called GrowthX, and it exists to be cited by AI answer engines. Its Collibra page opens like this:
Collibra doesn’t publish pricing on its website, so we went looking for it elsewhere. By cross-referencing UK government procurement filings, AWS Marketplace listings, and verified marketplace transaction data, we pieced together the real numbers…
checkthat.ai
It covers Anaplan. It covers Domo. It covers Coupa. Its Anaplan page opens by telling the reader that Anaplan doesn’t publish pricing, then hands them a per-user figure it attributes to a State of Nebraska government contract, and a median it credits to Vendr.
It does not mention Anaplan’s AWS listing at all.
So the number a buyer is most likely to encounter for Anaplan isn’t the one Anaplan filed on AWS. It’s the one an SEO agency assembled from government contracts, presented on a page engineered to rank, with Anaplan’s logo at the top. When I went to check that government figure against the source, the Nebraska contract wouldn’t render as readable text through any tool I tried, and the only places that quote a dollar amount from it are the same content farms. The most-cited number for what Anaplan costs traces back to a document almost nobody can open.
The public record of what your software costs is mostly other people’s SEO. Which, if you’ve read our work on why original data gets cited far more than opinion, is exactly the outcome you’d predict. Somebody did the work. It wasn’t the vendor.
Why this is about to get worse
Two pieces of research did the underlying work here, neither of them mine.
VisibleIQ classified 2,020 citations across four AI engines. For decision-stage questions, the ones that sound like “is X worth it” or “how much does X cost,” 93.7% of citations on Perplexity, Gemini and Claude pointed at third-party pages. Not the vendor. Agency blogs, comparison sites, independent reviews. We’ve seen the same divergence ourselves when we asked four engines the same buying question and got four different shortlists.
Then Suganthan Mohanadasan did something clever: instead of reading what ChatGPT said, he read the network traffic underneath it. He caught the model’s own saved reasoning while it tried to price two products. It went to the vendor pricing pages first, which is what you’d want. Then:
the pricing isn’t showing up directly in the search result, possibly because it’s loaded with JavaScript
ChatGPT’s saved reasoning, captured by Suganthan Mohanadasan
And then it gave up and wrote:
I can quote third-party sources since the official page is hard to parse and doesn’t show prices
ChatGPT’s saved reasoning, captured by Suganthan Mohanadasan
and noted it should “use citations from G2 where appropriate.”
His summary is better than anything I’d write: “A JavaScript pricing table doesn’t just rank badly, it hands your numbers to G2.”
That should worry you more than a ranking drop, because G2 and Trustpilot already function as citation multipliers for brands that never asked them to.
His case was a price the vendor published but rendered badly. The vendors in this piece are a harder version of the same thing. Their pricing pages don’t have a slow number or a hidden number. They have no number. The engine probes the vendor’s own domain, finds nothing to read, and falls back to whoever did answer the question.
Which, increasingly, is a content farm.
What this actually costs you
If you’re at one of the sixteen, the situation is this. Your price is public. You published a lot of it yourself. You just published it somewhere your buyers don’t look and your marketing team has never opened.
Meanwhile the answer your buyers get comes from a site you don’t control, quoting a number you didn’t choose, sourced from a government contract you signed in a different context, with no date on it.
Go back to where this started, because Braze and checkthat.ai are doing the same thing.
Braze marked that FAQ up as structured data, so “does Braze have public pricing: no” is machine-readable and ready to be lifted straight into an answer. GrowthX wrote a page engineered to be cited by AI engines answering the question “what does Anaplan cost.” Both are optimising for the same robots. Both will get read. The schema choice itself shapes which one gets surfaced.
The difference is that one of them answered with a number.
When a buyer asks what you cost, something is going to answer. If your contribution to that moment is a machine-readable “no,” you haven’t opted out of the conversation. You’ve just left the only useful answer to somebody who’s selling something else.
“Contact sales” didn’t keep your pricing private. It just meant you weren’t the one who got to explain it.
What to do about it this week
None of this requires publishing a rate card. It requires knowing what’s already out there with your name on it.
- Open your own AWS Marketplace listing. Search your brand on AWS Marketplace while logged out. If there’s a number there, that number is your public price, whatever your website says.
- Diff your regional pages. Rippling’s US page and UK page disagree. So might yours. Check every locale you serve, and check the currency actually matches the country.
- Search your brand plus “pricing” and read the top ten results. Somebody has written your pricing page. Find out who, what number they used, and where they got it.
- Check the government schedules. GSA Advantage, UK G-Cloud, state contracts. If you sell to the public sector, you have already published a rate.
- Ask the engines directly. Type “how much does [your brand] cost” into ChatGPT, Perplexity, Google AI Mode and Copilot. Our free brand visibility audit walks through the method if you want something repeatable.
- Decide who owns the answer. Right now, at most of these companies, nobody does. That’s the actual bug. Everything else is a symptom.
If what you find is wrong, fixing what AI says about your brand is its own discipline, and it starts with the entity data rather than the pricing page.
We publish what our own work costs, and we publish what the rest of the market charges, including the agencies we compete with. It isn’t a moral position. It’s that the question gets answered either way, and we’d rather be the ones answering it. If you’re selling B2B SaaS into AI-mediated research, the same logic applies to you, and the wider citation data says it’s getting more true, not less.
Method
I checked 21 enterprise SaaS vendors whose own pricing pages publish no number and gate pricing behind a contact form. The list was built from the top 5 brands by US organic traffic across 20 SaaS categories, using DataForSEO’s Bulk Traffic Estimation, then filtered to those with no published price. Selection was mechanical, not hand-picked.
For each, I searched AWS, Azure and GCP Marketplace, GSA Advantage, UK G-Cloud, US state contracts, purchase-data aggregators, review sites, regional variants of the vendor’s own site, and archived snapshots.
Every price here is verified. I read each one straight off the vendor’s own page, not a search snippet, and if a snippet was all I could find, I left it out.
The Braze FAQ answer is the one in the structured data on the Braze pricing article, not a paraphrase a search engine served up. The Rippling figures come from the pages themselves, whose page language is set to en-GB and en-IE. The Oracle figure comes from Oracle’s own PDF, which scrambles its price columns the moment you copy text out of it, so I lined each number back up with its product by hand.
Some things I want to be upfront about:
- Sixteen of twenty-one isn’t sixteen of the market. It’s sixteen of a small panel chosen from one traffic ranking. It tells you the pattern is common in this group. It doesn’t tell you the rate across all enterprise software.
- AWS list prices are not what anyone pays. The Anaplan gap makes that obvious. The claim here isn’t that anyone writes a cheque for the list price. It’s that a public number exists with the vendor’s name on it, and the vendor mostly doesn’t know.
- Five of the twenty-one publish nothing anywhere I could find: ADP, Paychex, Paylocity, Greenhouse and Lever. Their pricing pages have said the same thing for years. ADP and Paychex both own separate brands that do publish prices, Roll by ADP and SurePayroll, which are different products from the flagship and I’ve counted them as such.
- Five more only have prices through resellers or in archived pages, so I’ve kept them out of the count: Splunk, MuleSoft, SAP, Recruitee and Adobe Workfront. Splunk is the interesting one. Its own AWS listing shows no price, just “Request a private offer,” while a reseller publishes Splunk’s full per-GB rate table on the same marketplace.
- Four vendors were dropped after their pricing turned out to be real: Qlik and MicroStrategy had genuine self-serve pricing. CrowdStrike was pulled once its page showed actual figures. TriNet was cut because its pricing page contains a worked example with a number in it. The panel wasn’t built toward the answer.
- Government contract lookups produce false positives. Two GSA hits, one under ADP’s name and one under Splunk’s, both resolve to the same $82.30/hr generic staffing rate that has nothing to do with either company. A UK G-Cloud listing showing “£370 to £850 a unit a day” for Domo is a reseller’s consulting day rate, not a licence fee. Both would look like prices in a table.
- Prices move and pages change. Every figure here was checked in July 2026. AWS Marketplace listings can be edited without notice and without a changelog. If you’re reading this later and a number has moved, that is worth knowing and is itself the point.