Sits above Microsoft Fabric or AWS
Your platform holds the data. Your department heads hold the decisions.
Finance knows what "revenue" means. Operations knows which depot figures are estimates. Lake On Rails is browser software that puts each decision to the person who can answer it, records it against their name, and shows you whether it is still holding a year on.
It never moves or reads the contents of your data. The rules arrive already drafted, so your department heads approve an operating model rather than writing one — above Microsoft Fabric or AWS, or ahead of choosing either.
Ten questions, two minutes, and an honest answer that may be "a spreadsheet is still enough". Or skip it and talk to a person: the first conversation is free.
Revenue (Group)
Depot Utilisation
Vehicle Telematics Events
Inspection Records
Thirty of them have nobody's name against them. Those are the rows this exists to close.
What Lake On Rails is
One place where the decisions about your data are made, recorded and kept current
It is for businesses building a data platform, or already running one, where the decisions have outgrown the one person who used to hold them: which numbers are trusted, who is accountable for each dataset and report, what the rules are, how long things are kept, and whether any of it is actually being followed.
Those answers do not belong to your data team. The finance director defines what revenue means; the quality manager decides how long inspection records are kept. This is built so those people can make and record their own decisions in a browser, in a minute, in plain English — and so you can show afterwards that they did.
The real product, with sample data. A single score, six things that make it up, and a link from each to the screen where you fix it. See it full size
Where the line falls
Your cloud platform answers half of this
What Fabric and AWS answer
Where your data lives and how it is processed. Storage, pipelines, permissions, a catalogue.
Solved. Nothing here to pay for twice
What they don't
Who owns it, what the rules are, and whether any of it is actually happening.
That is the part Lake On Rails holds
Why a catalogue cannot answer itDoes any of this sound familiar?
Ten signs it is already costing you
None of these is a technology problem. Each one is a decision nobody was asked to make: who owns a number, which source is current, what happens when the person who knows is away. Tick the ones you recognise. Nothing leaves your browser.
Your count
0 of ten
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
This measures symptoms, not maturity, and it is not a benchmark against anyone. It is a mirror.
Why the order matters
Decide how you'll run your data before you decide where it lives
Whether you choose Fabric, AWS or nothing yet, the same decisions have to be made: who owns this, what it is for, how long you keep it. Make them now and the people who know the answers are still in the room. Make them after a migration and you are asking an implementation partner to guess, then living with the guess on everything they moved.
If you have not chosen a platform yet, this makes the choice easier. If you have, it makes the build smaller. If you are already live, it is what stops the drift.
Fabric or AWS? What survives the decision either wayTrue whichever platform you pick
- Who owns each dataset
- What your trust tiers mean
- What counts as sensitive, and how long you keep things
- Who approves access, and what "revenue" means
- Who does what when it breaks
This is the list nobody is usually working on. It is what Lake On Rails holds.
Why this, and not the catalogue you already have
A catalogue is built for the people who move the data. The decisions belong to the people who use it.
Microsoft Fabric and AWS give you a good catalogue: schemas, lineage, permissions, refresh times. All of it is written for the two or three people in your business who work with data for a living. None of it asks your finance director what revenue means, or your quality manager how long inspection records have to be kept — and those are the answers that decide whether anything in the platform can be trusted.
Schema41 columns
Lineage6 upstream
Permissions14 groups
Who owns the definition—
Which version is current—
Whether the rule was followed—
Illustration
The platform is not withholding this
It is not a gap in the product. A catalogue can record the shape of a table; it cannot record a judgement nobody has made yet. Accountability, a promotion rule, a retention period — those come out of a person's head, not out of a pipeline.
revenue_group
revenue_group_v2
revenue_group_final
Illustration
Nobody chooses a swamp
It is what a platform turns into when no one is empowered to say no to a source. Every source gets copied in, three versions of the same table appear with no rule about which is current, and people go back to their own extracts. The symptoms, in full.
Maturity score
Two datasets landed last month with nobody's name against them. Sent to Priya Shah to assign.
Illustration
And a sponsor can see it holding
This is what lets a senior leader roll it out rather than hope: one number that moves, a list of who has been appointed and what they have approved, and an alert when something lands with no owner. Drift shows up as a falling score, months before it shows up as an argument in a board meeting.
Who the owners actually are
The people who know what the data is for do not work in IT
Finance knows what revenue means. Operations knows which depot figures are estimates. Quality knows what must be kept and for how long. They run departments. They will not open a repository, write a query or raise a ticket to record that retention on inspection records is seven years, not three.
So every decision the product asks of an owner is takeable in a browser, in a minute, without any special vocabulary. It reaches them when they are appointed, rather than waiting for them to come looking. And because each decision is recorded against a name and a date, the sponsor above them can see it was made.
The five decisions, and whose they are
| Finance | What revenue means, and which cut the board sees |
|---|---|
| Operations | Which depot figures are real and which are estimates |
| Commercial | What counts as an active contract |
| Quality | Which inspection records have to be kept, and for how long |
| HR | What may never leave the building |
Not one of these is a question the platform team can answer for them.
What you are actually buying
What a data operating model records, and why you don't start from a blank system
Every tool in this category ships you an empty template. That is the weeks nobody has. Here, your one analyst doesn't author an operating model. They approve one, and change the bits that are wrong for you.
Tom Vance
Aisha Khan
Kept 36 months
Illustration
Everything the analyst knows, written down
A register of every dataset and report, with an owner and a steward by name, where it comes from, how sensitive it is and how long you keep it. This is the record that answers "what happens when she leaves?"
See the real screen
Register the dataset and name an owner
Set the classification and how long it is kept
Run the quality checks before it is promoted
Illustration
Seven procedures, written and ready to edit
Onboarding a dataset, granting access, promoting between tiers, handling a schema change and more. Each with a named role per step and a version history. Read one in full, here, now. If it is wrong for your business you will know inside a minute.
See the real screen
Maturity score
Illustration
A number your board can follow
One maturity score, the six things that make it up, and a one-page print for the board pack. It goes down when something slips, which is the point.
See the real screen
Before you ask
What we won't do
Nobody lies to make themselves look worse, so these are the most reliable sentences on the page.
- Move, transform or store your data. Where it connects to a platform it reads technical metadata only. Never the contents.
- Sell you a blank template and call it a product. The procedures, the tier criteria and the responsibility matrix arrive drafted.
- Take your money if nobody can carry half a day a week. That is the real cost, and for most businesses your size it is bigger than our invoice. What it asks of your people, in full.
- Tell you it fits when you have three reports and one analyst. A spreadsheet is honestly enough, and we will say so.
- Tell you which cloud to pick. We do not need to know which to be useful, and a vendor with a platform opinion is a vendor with a partner to feed.
- Charge you to leave. The register, the procedures and the audit trail export in open formats whether or not you stay.
Why I built this
One analyst. Nine definitions of "revenue".
I'm Andrew, the founder of Scorchsoft. We have built software for other people's businesses since 2010, and the same conversation kept happening at handover of every data platform: "who owns this number now?" Nobody had written it down. Every report had its own model, because reusing someone else's was harder than starting fresh, and the definitions lived inside the files and inside one person's head.
The tools that fix that were built for companies with a governance team and a six-figure budget. Our customers had one to five people and a spreadsheet. So we turned the operating model we kept writing in Word documents into a product, with the procedures already in it, and priced it for the business we actually work with.
It is new, and I would rather you knew that from me than found it out in month four. If you write to us, a person replies, and on anything that matters that person is me.
Andrew, founder, Scorchsoft Ltd · info@scorchsoft.com
What Scorchsoft's history does and doesn't prove
Sixteen years of custom software, fifty-plus enterprise clients, a 4.9 Google rating and Cyber Essentials certification are facts about Scorchsoft. They do not make Lake On Rails proven. They do mean the company behind it has outlived three recessions and will answer the phone in year four, which is the thing a new product cannot otherwise promise.
We have deliberately not put Scorchsoft's client logos on this page. They are for custom software, and a sharp finance director would notice.
About ScorchsoftThree ways to work with us
Which of these questions are you being asked right now?
Each plan answers one. Start there; you can move up when the question changes.
Foundations
"Which number is right?"
£8,000 / year
- Dataset register with named owners
- Trust tiers and classification
- Roles and responsibility matrix
- The seven procedures
- One platform connection, up to 10 people
Professional
"What happens when she's on holiday?"
£18,000 / year
- Everything in Foundations
- Approval workflows and notifications
- Training and learning paths
- Maturity roadmap and health scoring
- Audit trail exports, API and MCP server
- Several platform connections, up to 50 people
Enterprise
"Can you show me?"
£30,000 / year
- Everything in Professional
- Single sign-on
- Compliance reporting and evidence packs
- Custom workflow templates
- Unlimited people and connections
Prices exclude VAT. Monthly payment available. Consulting is priced separately.
Being straight with you
We are looking for our first ten
Lake On Rails has no customer references yet, and we are not going to pretend otherwise. You would be early. Here is what early gets you.
- The founder on your discovery, and a named line to the people building it, for as long as you are a customer.
- A say in what comes next. The first ten set the roadmap, because the first ten are the roadmap.
- Your operating model in open formats, whether you stay or not. It is the product of the work, not a feature you have to keep paying to export.
- An honest "not yet" in the first conversation if it does not fit. We would rather lose the deal than the reference.
In return we ask for a monthly half-hour call for the first year, straight feedback, and, only if it works, a reference.
This fits you if
It doesn't, yet, if
Where you are, and what the next step costs
Five rungs, and the first two are free
- 1
Count the signs
Two minutes, above. Nothing leaves your browser.
Free
- 2
Forty-five minutes with us
You, whoever runs reporting, and an honest answer.
Free
- 3
Discovery
A register of what you have, an owner per line, your tiers defined. Yours to keep.
Fixed fee
- 4
The platform
Foundations, Professional or Enterprise, with your model already in it.
From £8,000 a year
- 5
The quarterly review
We run it with you and report to your sponsor. The thing that stops the drift.
Optional
Questions your team will ask
Straight answers
If yours is not here, ask us. You will get the same kind of answer.
All the questionsWe don't have a data lake. Is this for us?
Yes, and that is the most common situation. Most of the businesses this was built for run Power BI or similar on their application databases and are weighing Microsoft Fabric against AWS. The decisions recorded here — who owns each dataset, what counts as trusted, what is sensitive, how long you keep things — are true whichever platform you choose, and they are the input that decision needs.
How is this different from what Fabric or AWS already gives us?
The platform answers where your data lives and how it is processed. It does not record who is accountable for a dataset, what your promotion rules are, which written procedure applies, or whether any of that is actually being done. Lake On Rails is that layer. It adds nothing to the rows the platform already covers, so you do not pay twice. More on Fabric and AWS.
Do we need a data governance team to use it?
No. It is built for organisations with one to five people who touch data, none of them full-time on this. What it does need is one person who can carry roughly half a day a week, and a sponsor who cares. If nobody will carry it, it will not work, and we would rather say so up front.
Does it move or store our data?
No. It never moves, transforms or stores the contents of your data. Where you connect it to a platform it reads technical metadata only: dataset names, schemas, owners and refresh times. You can also run it entirely on metadata you enter yourself, with no platform connection at all.
What does it cost?
Three annual plans at £8,000, £18,000 and £30,000 a year, excluding VAT, each a 12-month commitment and all sold with a person on the other end. The first two can be paid monthly instead, which costs a little more over the year and is shown as a total rather than hidden; Enterprise is annual only. Consulting is priced separately and is optional. There is no free tier. The first conversation is free. Pricing in full.
Built in Birmingham by Scorchsoft: custom software for other people's businesses since 2010, Cyber Essentials certified, rated 4.9 on Google across 30+ reviews. Those are facts about Scorchsoft, not proof of this product. Why that matters.
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.