Why "Refinitiv alternatives" is the wrong shopping list
Refinitiv is not a single product, so there is no single replacement, and the only useful first question is which of its layers your invoice is actually paying for. Most institutions license some mix of four distinct things, and each has a different competitive set. Treating them as one purchase is how a firm ends up replacing a terminal it needed and keeping a data licence it does not query.
The four layers, in the order they usually appear on an enterprise contract:
- Desktop seats. Workspace, and before it Eikon: a licensed human sitting in front of a screen.
- Real-time market data feeds. Consolidated venue data, entitlements, redistribution rights, non-display permissions.
- Reference, identifier and entity data. Security master, cross-identifier mapping, corporate actions, issuer and legal entity hierarchies.
- Index and ETF constituent data. The holdings of each index and fund, dated, weighted, and mapped to identifiers.
The naming history makes procurement harder than it should be. Refinitiv is the former Thomson Reuters financial and risk business, and it now sits inside the London Stock Exchange Group, which is why the pages that rank for this query label it inconsistently: OpenBankingTracker's list, fetched on 17 September 2026, titles it "Refinitiv (LSEG) Alternatives," while Wall Street Prep's comparison still refers to the product as Refinitiv Eikon. If you are still forming a view of the platform itself before shopping, our Refinitiv Review 2026 covers it layer by layer rather than as one monolith.
One more consequence of the layer structure deserves saying plainly. You can cancel one layer without touching the others, and the layer that is easiest to unbundle is almost never the terminal.
What do the ranking alternatives lists actually compare?
The three pages that currently rank for this query each compare a real thing, but none of them compares the same thing, and none prices the constituent-data layer at all. Knowing which list you have already read tells you which layer you have already shopped for.
Here is what each covers, based on the pages as fetched on 17 September 2026:
- Datarade's competitors page lists Exchange Data International, Cbonds, Coresignal and QuantLens as the top competitors and alternatives to Refinitiv, and frames the choice around finding a stock market data provider. It is a marketplace view: useful for discovering specialist data suppliers you had not heard of, which is genuinely valuable, and it does that better than a terminal comparison ever will.
- OpenBankingTracker's list opens by defining an alternative as "another bank-technology vendor to compare for APIs, go-to-market story, and customer proof," and says it lists ten peers so you can shortlist without building a market map. That framing is honest about what it is: a vendor landscape for banking technology and risk analytics buyers, not a data-licence comparison.
- Wall Street Prep's comparison covers Bloomberg, Capital IQ, FactSet and Refinitiv Eikon, with a dedicated pricing section and a "who is this best for" section for each. It is the most substantial of the three at roughly six thousand words, and its strength is workflow fit for banking and investing analysts.
Review-aggregator category pages also rank for this query. They were not part of the material reviewed for this article, so treat any figure on them as secondhand until you match it to vendor documentation or a quote.
The shared gap is the same in all three: they compare vendors, not layers. A page that puts a marketplace data supplier and a desktop workstation in one ranked list is answering two different purchase decisions with one ordered list, and the index and ETF constituent line item does not appear in any of them.
Layer 1: replacing the Workspace terminal seat
A desktop seat is worth replacing when the person holding it uses fewer than a handful of its functions, and worth keeping when their workflow is search, screening and ad-hoc analysis under time pressure. Start by separating users who chose the terminal from users who inherited one during onboarding.
The credible alternatives in this layer are other workstations, because the thing being replaced is an interface, not a file.
Bloomberg Terminal. How it differs is a question for Bloomberg's own product documentation: nothing fetched for this article dates a claim about its positioning, so treat characterisations you read elsewhere, including the common one about its messaging network, as unverified until you check the source. Fits desks whose counterparty communication already lives there. What you give up switching to it is nothing on the data side that this article can verify from sources fetched today; Wall Street Prep's page includes both a "What is Bloomberg Best For?" section and a "You could probably live without Bloomberg if…" section, which is a fair way to frame it. Not for teams whose seat usage is monthly rather than daily.
FactSet Workstation. This article has no dated source for how FactSet positions the workstation relative to the incumbent, so read the current scope from FactSet's own product pages rather than from a comparison post. Fits research and portfolio teams who want their analytics and their data in one place. What you give up is any workflow your team built around Refinitiv-specific functions and Excel formulas, which is real switching cost measured in analyst weeks. Not for a desk that needs a broad multi-asset real-time interface and treats analytics as secondary. A layer-by-layer read is in FactSet vs Refinitiv: 2026 Comparison, and the broader field in FactSet Alternatives: 2026 Options.
S&P Capital IQ. The product's current scope is published by S&P Global, and nothing fetched for this article dates a claim about where its coverage is strongest, so verify the fit against their own pages. Fits corporate finance, M&A and credit workflows. What you give up is any dependence you had on the incumbent's real-time market depth inside the same window. Not for trading desks where the screen is a live market monitor.
Research-search platforms in the AlphaSense mould. Different in that they replace one job the terminal does, document and transcript search, rather than the whole seat. Fits analysts whose actual terminal usage is reading filings and transcripts. What you give up is everything else the seat did. Not for anyone who also needs pricing, screening or reference data from the same tool.
One caution on this layer specifically. Per-seat figures circulating on comparison pages are not vendor-published rate cards, and LSEG does not publish one for Workspace. Verify any number you plan to budget against with a written quote.
Layer 2: real-time market data and the feed contract
The real-time feed is the hardest layer to displace, and this article is not going to pretend otherwise. A consolidated real-time licence is not just data: it is exchange redistribution agreements, per-user entitlement enforcement, non-display permissions, latency commitments and an integration into your ticker plant that took someone a quarter to build.
Four things make this layer sticky in ways a feature comparison never captures:
- Exchange agreements. The venue, not the vendor, sets the terms of redistribution, and every new supplier means new paperwork and new reporting obligations.
- Entitlement enforcement. Someone has to attest who is entitled to what, per user, per venue, per display and non-display use. That machinery is part of what you are buying.
- Latency and normalisation. Cross-venue symbol normalisation is boring, expensive work, and an incumbent has already done it for your universe.
- Downstream integration. Order management, risk and pricing systems read the feed in its existing shape.
The more common finding in an audit is not that the feed is replaceable but that it is over-specified. Teams whose models run on end-of-day closes and dated holdings frequently hold real-time entitlements they query only incidentally, because the entitlement came bundled with a seat or a historical package. That is a downgrade conversation, not a replacement conversation, and it is usually the cheapest win in the whole exercise.
For consolidated real-time coverage across global venues, incumbent vendors are genuinely hard to beat. If that is the layer you are shopping, the honest answer is to negotiate rather than to migrate.
Layer 3: reference, identifier and entity data
Reference data is the layer that answers "what is this instrument and who issued it": security master records, cross-identifier mappings, corporate actions history, and issuer to legal entity hierarchies. It is the layer that silently determines whether two datasets join correctly, which is why it is the worst place to improvise.
The identifier question is where cost and licensing meet, and the four common identifiers do not come from the same place or under the same terms:
| Identifier | Issued by | Practical licensing note |
|---|---|---|
| FIGI | Bloomberg's Open Symbology / OpenFIGI | Published as an open standard; check current terms at the source |
| ISIN | National numbering agencies, coordinated through ANNA | Terms vary by numbering agency and jurisdiction |
| CUSIP | CUSIP Global Services | Licensed; commercial redistribution terms apply |
| SEDOL | London Stock Exchange | Licensed; terms set by the issuer |
Do not take licensing terms for any of these from a blog post, including this one. Read the issuing body's current terms, because the commercial conditions on identifiers change and a redistribution clause you assumed is the kind of mistake that surfaces in an audit rather than in a procurement meeting.
Where this touches the constituent layer is mapping. AmericanETP includes Bloomberg FIGIs in the holdings files of its constituent lists, which removes one mapping step for a team joining ETF holdings to an internal security master keyed on FIGI. That does not replace a reference data licence, and it is not sold as one; it removes a join that otherwise has to be maintained. The layer-level differences are set out in AmericanETP vs Refinitiv (LSEG) 2026, and the Bloomberg extensions page covers the identifier side in more detail.
Not for firms that need full global corporate actions and entity hierarchy coverage: that is a reference data purchase, and no constituent feed substitutes for it.
Layer 4: index and ETF constituent data, the line item nobody unbundles
Index and ETF constituent data is the dated list of what each index or fund holds, with weights and identifiers, refreshed as the composition changes. It is the layer that none of the ranking alternatives pages prices, and it is the one most often bundled invisibly into an enterprise contract alongside seats and feeds.
There is a simple test for whether you are paying twice. Ask three questions inside your own firm:
- Which systems read constituent files today, and which vendor do those files come from?
- Is there a separate line item for index or constituent data on the contract, or is it folded into a package?
- If you cancelled the package tomorrow, which of your models would stop running, and would any of them stop for a reason unrelated to constituents?
If the answer to the third question is "the backtests need holdings, nothing else," you have found an unbundlable line item.
AmericanETP exists to be that line item and nothing more. Its shape, stated factually: constituent lists update twice daily, with a primary run at 6pm EST and a secondary run at roughly noon EST; coverage spans 3878 US and global indexes and US-traded ETFs; a fundamentals file carries the ETF description, leverage and constituent counts; archived constituent datafiles go back to July 2009; and delivery includes FTP access alongside the downloads area, with a constituent change report for composition changes.
Pricing is published rather than quoted: $1500 per year or $150 per month for an individual, and $2500 per year or $250 per month firm-wide. The company was founded in the 1990s as MasterDATA, predating most of the ETF data vendors it now sits beside on comparison pages, and was renamed to AmericanETP to avoid confusion with IBM's use of the earlier term.
Not for a firm that needs intraday constituent snapshots, real-time NAV, or a single vendor for every layer on one contract. If consolidation onto one supplier is the mandate, a specialist file feed adds a vendor rather than removing one. For how the constituent-only category compares across suppliers, see ETF Global Alternatives: 2026 Options.
How much does Refinitiv cost, and why alternatives lists avoid the number
LSEG does not publish a public rate card for Workspace seats or enterprise data licences, so any specific per-seat figure on a comparison page is secondhand unless it is attributed to LSEG's own materials. Wall Street Prep's page does include pricing sections for each platform it covers, which is more than most roundups attempt, but a figure without a vendor citation is a starting point for a conversation, not a budget input.
What you can verify sits in your own contract. These are the cost drivers worth reading before you shop anything:
- Seat counts and seat types, including inactive seats nobody has cancelled.
- Real-time entitlements by venue, and whether anyone queries them.
- Redistribution rights, which govern whether data can move to a downstream system or client.
- Non-display use, which is priced separately from human display use at most vendors and is the clause that catches quant teams.
- Derived data clauses, which determine what you may publish from a model trained or run on the data.
- Multi-year escalators, which set your renewal number before you negotiate it.
This is why published flat pricing changes the character of the purchase rather than just the amount. A figure you can read on a page is a figure finance can model for three years without a quote cycle, which is the actual argument for unbundling a small layer even when the savings are modest. Our breakdown of the cost drivers is in Refinitiv Pricing: What It Costs in 2026.
A four-step process for running the replacement
Run the audit before the shortlist. Almost every failed migration fails because the firm bought a replacement for a layer it misidentified, and the fix costs more than the original licence.
1. Inventory by layer over 90 days. Log what users actually open and what systems actually pull, tagged to one of the four layers. Ninety days catches month-end and quarter-end index rebalances, which a 30-day sample misses entirely.
2. Separate display from non-display and derived use. Human screen use, machine use inside a model, and anything you publish externally are three different licensing positions. Getting this wrong in either direction is expensive: over-licensing wastes money, under-licensing is a compliance event.
3. Run a parallel file comparison on a fixed universe before cancelling anything. Pick 50 to 100 indexes and ETFs that matter to your models, pull both sources for the same as-of dates, and reconcile. Sixty days of parallel running costs one month of a specialist subscription and tells you more than any trial.
4. Renegotiate the remaining contract with the unbundled items removed. Go into renewal with a documented list of what you are dropping and what you are keeping. A vendor cannot price a bundle you have already decomposed.
The reconciliation test in step three is the one that matters, and it has two parts: do the constituent counts match on the same as-of date, and do the weights sum and align within tolerance. Legitimate differences have explanations you can name: different as-of conventions around rebalance dates, cash and derivative line handling, share-class treatment, fractional-share rounding, and whether a pending index addition is reflected before or after the effective date. Unexplained differences in membership itself, a holding present in one file and absent in the other with no corporate action behind it, are errors, and you should resolve them before cancelling anything.
Which alternative fits which institution
The right first move depends on which layer carries your cost and which carries your risk. For several buyer types, the correct answer includes keeping the incumbent.
| Buyer type | Attack this layer first | Realistic outcome |
|---|---|---|
| Quant funds and backtesting teams | Layer 4, constituent data | Historical and daily holdings move to a specialist file feed with FIGIs and archives; real-time entitlements often downgraded rather than cancelled |
| RIAs | Layer 1, seats | Fewer terminal seats, one shared research tool, constituent data bought as a flat-priced line item |
| Banks | Layers 2 and 3 | Feeds and reference data stay with the incumbent; savings come from entitlement cleanup and seat rationalisation, not migration |
| Securities exchanges | Layer 3, identifiers and reference | Identifier licensing reviewed at the issuing body; incumbent retained where redistribution rights are already papered |
| Data vendors and fintech teams building on feeds | Layer 4, then Layer 3 | Constituent and holdings data sourced directly with redistribution terms clarified upfront; terminal seats usually irrelevant |
| Large institutional investors | Layer 1, then Layer 4 | Seat audit first because seats are the largest headcount-linked cost; constituent data unbundled second |
If your requirement is broad fund and manager research rather than index membership, the competitive set is different again, and Morningstar Alternatives: 2026 Options is the more relevant comparison.
What the incumbent still does best is breadth under one contract: consolidated real-time coverage, reference data, and a desktop workflow procured, papered and supported as one relationship. If your constraint is vendor count rather than cost, or if your traders live in the terminal all day, staying is the correct decision and the audit will tell you so.
FAQ
Is Refinitiv as good as Bloomberg?
They are strongest at different things, and the comparison only resolves per layer. Wall Street Prep's comparison, as fetched on 17 September 2026, organises its assessment around "what is this best for" and "you could probably live without this if" sections for each platform rather than declaring a winner, which reflects how the decision actually works. For a trading desk whose counterparty communication already runs through one platform, that dependency is the deciding factor and no data comparison overrides it.
Is Thomson Reuters the same as Refinitiv?
No. Refinitiv is the former financial and risk business of Thomson Reuters, separated out and rebranded, while Thomson Reuters continues as a separate company in legal, tax and news. Procurement documents from different years may use either name for the same subscription, which is worth checking when you reconcile old contracts.
Is Refinitiv now called LSEG?
Refinitiv sits within the London Stock Exchange Group, and the product line is marketed under LSEG branding, with Workspace as the current desktop product name. Third-party pages continue to use both labels: OpenBankingTracker's page, fetched 17 September 2026, titles its list "Refinitiv (LSEG) Alternatives," while other ranking pages still say Refinitiv Eikon. Assume any list you read may be describing an older product generation.
How expensive is Refinitiv?
There is no public rate card for Workspace seats or enterprise data licences, so a firm figure requires a quote. The cost drivers you can verify yourself are seat count and seat type, real-time entitlements by venue, redistribution and non-display rights, derived data clauses, and multi-year escalators. Any specific number you see on a comparison page should be treated as secondhand unless it is attributed to LSEG's own published materials.
What are the best FactSet alternatives if I am already leaving Refinitiv?
That depends on which layer you were using FactSet for, the same as with the incumbent. If the requirement is portfolio analytics inside a desktop workflow, the comparison set is other workstations; if it is index and ETF holdings feeding a model, it is specialist file feeds. FactSet Alternatives: 2026 Options works through both cases separately.
Can I replace only the ETF and index constituent part of a Refinitiv contract?
Frequently yes, and it is the most commonly overlooked unbundling. Constituent data is a dated file, not an interface, which makes it easy to reconcile against a second source and easy to cut over once the reconciliation passes. Run a parallel comparison on a fixed universe of indexes and ETFs for 60 days, confirm the counts and weights reconcile on matching as-of dates, then remove the line item at renewal.
Constituent lists for 3878 US and global indexes and US-traded ETFs, updated twice daily with archives back to July 2009, at published pricing you can budget against: see what AmericanETP delivers, or get in touch to set up a parallel comparison against your current source.