FactSet vs Bloomberg Cost per Seat for Small Firms

Neither vendor publishes a price, so on the face of it this is a comparison between two unknowns. It is not, and the reason is easy to miss. FactSet Research Systems is listed on the New York Stock Exchange and files its total subscription value and its user count with the SEC every quarter. Bloomberg L.P. is private and files nothing. So one of these two prices can be bounded from audited documents and the other cannot be bounded at all, which changes how much weight you should give every figure you are about to read.

Here is the short version for a firm with two to ten seats. A Bloomberg Terminal is reported at roughly $24,000 to $32,000 a seat a year. FactSet's own filings put its average user at $10,027 of annual subscription value at May 31, 2026, and Vendr's median FactSet contract is $25,160 a year for a whole deal rather than one seat. So the realistic gap is around two and a half to three times per seat, not the twenty times a retail-priced roundup implies and not the modest difference the vendor-neutral comparison articles suggest. Below is where each of those numbers comes from and what it is actually measuring.

How much does FactSet cost compared with Bloomberg?

Every figure in this table was read at source on September 11, 2026, and the fourth column is the one that decides how much to trust the third.

Why one of these prices can be checked and the other cannot

This is the part no comparison article covers, and it is worth a minute because it tells you which numbers to argue with in a negotiation.

FactSet reports Annual Subscription Value, which it defines as the forward-looking revenues for the next 12 months from all subscription services currently supplied to clients. At May 31, 2026 that was $2,484.3 million, with 9,130 clients and 247,766 users. Divide the first by the third and the average FactSet user accounts for $10,027 a year. That figure is a ceiling on the average seat rather than a seat price, because ASV also includes data feeds, CUSIP Global Services and enterprise solutions that no named user sits behind. The true average workstation is lower.

What that ceiling does is settle an argument. If FactSet's average user really cost $12,000, the company would be reporting about $2,973 million of subscription value at that user count. It reported $2,484.3 million. So the $12,000 entry figure that circulates is above FactSet's own average, which does not make it wrong for a loaded workstation but does mean most FactSet users are not on one.

Run it back and the average is sliding: $11,447 in fiscal 2023, $10,518 in fiscal 2024, $10,136 in fiscal 2025, $10,027 at May 2026. The cause is in FactSet's own commentary. Users grew 12.4% in the year to May 2026 while subscription value grew 6.4%, and quarter after quarter FactSet names wealth management as the driver of user growth. Advisor seats are cheaper than institutional workstations, and they are diluting the average. The full series, with the arithmetic shown period by period, is on our FactSet cost page.

None of this is available for Bloomberg. There is no filing, no user count, no subscription value, and therefore no way to check whether $24,000, $31,980 or anything else is representative. The published Bloomberg figures disagree by around $8,000 a seat, and we traced the likely cause of part of that spread to renewal escalators in Bloomberg Terminal cost. It remains an argument between sources with no referee. Capital IQ sits between the two, a listed parent with no seat-level disclosure, and Capital IQ vs FactSet cost per seat for small firms works that comparison through. Morningstar Direct is the one other vendor that can be bounded the FactSet way, and with a cleaner denominator, on Morningstar Direct cost.

What does a small firm actually get for the difference?

The honest answer is that these platforms are strong in different places, and the per-seat gap is smaller than the capability gap in one direction and larger in the other.

That last row matters more than it looks for a small team, because it changes how many licenses you need rather than the rate. Three analysts sharing one office desk need one Open Terminal license. The same three wanting to log in from home need three Bloomberg Anywhere licenses. The mechanics are set out in Bloomberg's own manual and worked through on Bloomberg Anywhere subscription cost.

Which should a small investment firm buy?

Work from what your people open every day rather than from the brand.

Buy Bloomberg if the work is trading, fixed income or anything where the messaging network is how counterparties reach you. There is no substitute for Bloomberg chat, and firms that leave it usually come back for that reason alone rather than for the data. If your counterparties assume you are on it, the seat is a cost of doing business.

Buy FactSet if the work is research and portfolio analytics, especially if your models live in Excel or your own systems. The pipeline is where FactSet earns the seat, and at an average of $10,027 of subscription value per user it does so at roughly a third of a Terminal. Be specific on the call about the four or five data sets you would open daily, because the quote is assembled from that list rather than from a tier you pick off a page.

Buy neither, yet if what you actually need is company financials, estimates, screening, charting and client-ready output. That is the most common case at firms under about ten people, and platforms that publish their prices cover it. Koyfin's advisor tier is $209 a month billed annually, which is $2,508 a year, re-read at source on September 10, 2026. TIKR Ultimate is $79.95 a month billed annually, which is $959 a year, read on September 6, 2026. Those are a quarter and a tenth of FactSet's average subscription value per user. What they do not carry is portfolio analytics, risk models, fixed income depth or an institutional data pipeline, and if you need those, you are back to the first two paragraphs. The tier-by-tier detail is on Koyfin pricing, and the wider category is sorted in how much stock analysis software costs.

The work neither seat covers

Two gaps come up repeatedly at small firms, and both get discovered after a contract is signed.

The first is people and counterparties. A terminal tells you what a company reports. It does not assemble the public record on a management team, a fund's principals or a private counterparty you are about to transact with, and that job is not a market data problem at all. It is a public-source sweep that has to end in something sourced and timestamped so it can sit in a file, which is why firms that run this regularly use a tool that turns a public-record sweep into a sourced, timestamped report rather than a research subscription. Budget it separately, because no seat covers it.

The second is whether an idea has actually worked. Both platforms will show you a company's history and both will screen for a condition today. Neither will take a rule such as "hold while the price is above its 200-day average, otherwise hold cash", run it across two decades of daily history with trading costs charged, and tell you whether it beat buying and holding. That is a different kind of tool, and it is the one thing on this page you can get for less than the sales tax on a Terminal seat.

That narrow job is what AgentTrading does. You state a thesis in plain English, it restates that as an explicit rule and shows you the rule before anything runs, then tests it on 20-plus years of split and dividend adjusted daily data with a 0.1% cost per trade charged by default, plots it against buy and hold, and stamps a verdict that includes UNDERPERFORMED when the idea loses. It is research and analysis only, with no execution, no brokerage connection and no personalized advice, and it starts at $19 a month. For registered firms, backtesting software for financial advisors covers what the SEC marketing rule requires once a backtest reaches a client, and AI equity research covers what AI research platforms do and do not automate.

Other platforms sold the same way

FactSet and Bloomberg are not unusual. S&P Capital IQ publishes nothing either, and a per-user figure has to be reconstructed from real contracts, which is done in Capital IQ cost. YCharts publishes a full feature matrix and no price, reconstructed on YCharts pricing. If the reason you are pricing a terminal is that a Bloomberg seat is up for renewal, Bloomberg Terminal alternative for financial advisors prices the replacements at the tier a registered firm is actually permitted to buy, which is where most comparison articles go wrong. The broader research stack is on investment research software.

Sources: FactSet Research Systems Inc. Form 10-Q for the quarter ended May 31, 2026 and Forms 10-K for fiscal 2025, 2024 and 2023, filed under SEC CIK 0001013237 and read at sec.gov on September 11, 2026; factset.com/factset-pricing, rendered and checked September 11, 2026; vendr.com buyer guides for FactSet and Bloomberg and costbench.com, read September 11, 2026; koyfin.com read September 10, 2026 and tikr.com read September 6, 2026. The per-user and per-client figures for FactSet are our arithmetic on FactSet's reported numbers, not figures FactSet publishes, and FactSet discloses no average or typical contract value anywhere in its filings. Every Bloomberg figure here is a third-party estimate, because Bloomberg publishes none. No figure on this page is a quote, and a price either vendor gives you supersedes all of it.