How it compares

An operating model you can run with the people you have

There is no shortage of things you could buy. Spreadsheets and shared documents, the governance services in your own cloud, a mid-market catalogue, an enterprise platform, or a consultant who writes you a model and leaves you to keep it current. What Lake On Rails offers against all of them is a ready-made operating model for a small data team: ownership established, decisions run as approvals, and the evidence kept without anyone assembling it. Here is the honest version of that, including when you should choose something else.

What else you could do

All five of these are real choices, and for some businesses they are the right one. We would rather name them than pretend the market is a gap with us in it.

Option one

Spreadsheets, documents and the tools you already have

A workbook, a SharePoint folder, a Jira board, a Confluence page. Free at the point of use, and honestly enough for a handful of reports and one analyst. It stops working when nobody remembers to update it, which is usually the first time that analyst is away. Nothing chases an owner, nothing records an approval, and there is no number to show a board.

Choose it if: you have a handful of reports and no date on the calendar.

Option two

The governance services in your own cloud

Microsoft Purview and Entra ID alongside Fabric; Glue, Lake Formation and DataZone on AWS. Strong on discovery, lineage, permissions and platform audit, and priced separately from the platform itself — usually by consumption, so the bill grows with what you scan and store. They record a contact per asset rather than an accountable owner, and they do not hold a responsibility matrix, a written procedure with an approval, or a measure of whether any of it is being followed.

Keep it: Lake On Rails reads from it and adds nothing to the rows it already covers.

Option three

Mid-market catalogue and governance products

Dataedo, Secoda, OvalEdge, OpenMetadata and Collate, and others. Real products at real mid-market prices, mostly built around the catalogue: documenting assets, glossaries, lineage, discovery, and increasingly a workflow or two. Where they differ from this is what arrives on day one — they give you the system; the operating model, the procedures and the responsibility matrix are still yours to write.

Choose one if: cataloguing and discovery are the problem, and you have someone to write the rest.

Option four

Enterprise governance platforms

Collibra, Alation, Informatica, Atlan. Excellent at scale, built for hundreds of sources, thousands of users and a team to run them, and sold as quoted enterprise agreements rather than published prices. They are configurable enough to express almost any operating model, which is the point and also the work: what they express is the one you define.

Choose one if: you already have a data team and a governance lead, and the problem is scale rather than ownership.

Option five

A consultant-led model you maintain yourselves

A good consultancy writes you a framework, a RACI and a set of procedures, and hands them over. The thinking is often excellent. What decides whether it works is the year after: the documents live in a drive, nothing prompts anyone, and the only measure of whether they are followed is somebody's opinion at the next steering meeting.

Choose it if: you want the thinking and you already have somewhere disciplined to keep it running.

What we are

The operating model, arriving already written

A register with accountable owners, tier and retention rules, a responsibility matrix, seven procedures and the courses, all populated on day one and then run as approvals with a score that moves when they slip. It sits above whatever platform you choose, and it is priced for a business that has one to five people who touch data.

Choose it if: the gap is ownership and follow-through rather than cataloguing or scale.

Capability by capability

What each one actually does

"Yes" means built in. "Partial" means possible with configuration, or not really the tool's job. "Varies" means the products in that column differ enough that one answer would be wrong. Our own column is the only one we can speak for with certainty; the rest is our reading of published documentation, and our own column is honest about where the others are stronger.

Capability Lake On Rails Cloud native Mid-market catalogues Enterprise platforms Spreadsheet
Technical metadata
Dataset catalogue and schemaYesYesYesYesPartial
Automated metadata sync from the platformYesYesYesYesNo
Technical lineagePartial — read from the platform where it has itYesYesYesNo
Cost and capacity monitoringNo — not what this isYesNoPartialNo
The operating model
Named accountable owner per datasetYesPartial — one contact fieldYesYesPartial
Responsibility matrixYes, pre-filledNoPartial — roles, not a matrixPartialPartial
Written procedures, shippedYes — sevenNoNo — you write themNo — configurable, not suppliedNo
Procedure approval and version historyYesNoVariesPartialNo
Trust tiers with a promotion ruleYesPartialPartialYesNo
Running it
Approval workflows for access, onboarding, promotionYesPartialVariesYesNo
Maturity score and roadmapYesNoVariesPartialNo
Ownership-gap alertsYesPartialVariesPartialNo
Role-based training with completion trackingYesNoNoVariesNo
Board-ready one-page summaryYesNoVariesPartialNo
Append-only audit trail of decisionsYesPartial — platform events onlyVariesYesNo
Fit for a business your size
Runs without a governance teamYesPartialPartialNoYes
Useful before a platform is chosenYesNoPartialPartialYes
Works across Fabric and AWSYesNo — one cloud eachYesYesYes
How it is pricedPublished: £8,000–£30,000 a year, all inSeparately from the platform, mostly by consumptionSubscription, often per user or per assetQuoted enterprise agreementsFree, plus the analyst's time

Assessments of other products are our reading of their published documentation and of typical mid-market deployments, last reviewed in September 2026. They are not a benchmark, and they are not a substitute for asking a vendor directly. We deliberately do not publish other people's prices: they change, they are usually quoted rather than listed, and a stale number in our favour would be the least honest thing on this page. If we have a row wrong, tell us and we will correct it.

Being straight with you

Where the others are stronger

Cost control and platform monitoring

Tools built around one cloud — capacity attribution, workspace monitoring, cost per team — do that far better than a cross-platform product ever will. We do not compete there. If "save money on Fabric capacity" is the problem, buy that, and use Lake On Rails for who owns it and how it is run.

Lineage and column-level discovery at scale

Your platform's catalogue sees every table and every join. Lake On Rails reads what the platform exposes and holds the organisational layer above it. If you have hundreds of sources, the catalogue is the place for the technical detail.

Enterprise scale and a mature ecosystem

Collibra, Alation and Atlan have integrations, partners and reference customers we do not. If you are a thousand-user, hundred-source organisation with a governance function, you are their customer, not ours.

Cataloguing depth at mid-market prices

If what you actually need is to document what you have — glossaries, column-level documentation, discovery across a lot of sources — a catalogue product does that better and often for less. Some of them will also sit alongside this quite happily. The question to ask yourself is whether your problem is that nothing is written down, or that nobody is accountable for it.

Judgement about your business

A good consultant who knows your sector will make better calls about what your tiers should mean and who should own what than any product will. We ship a model that is right for most mid-market businesses, not one that is right for yours in every particular. Where the two disagree, take the consultant's answer and record it here.

The first step costs you nothing

Forty-five minutes with whoever runs your reporting

We tell you honestly whether this is worth doing at all, and roughly what it would take. If the answer is not yet, you will hear that. "Not for us" is a fine outcome, and a better one than a slow maybe.