Know who uses your data, what it costs, and whether access is under control.
Data Control Tower is a read-only view of your data estate that ranks what to fix and names the team that fixes it. We set it up, and go through the findings with you every quarter.
A free pilot for 1 company · Applications close November 20
0days00hours00min00sec
Every screen and video on this page shows Halvane Industries, a fictional company. Its people and numbers are made up.
Could you tell your CEO who used your data last month, what it cost, and who can still reach it?
Your catalog can’t tell you who read your data last week, what that cost, or that someone who left last month still pulled it at midnight. The answers are already on record. Nobody has the time to read them.
| What goes unanswered | What it costs |
|---|---|
| Is our data used? | You pay to store and secure tables nobody reads, and nobody knows which ones to retire. |
| What is it used for? | Each team sees its own corner. Nobody sees which departments use which data, or which data people combine by hand every week. |
| What does it cost? | The bill arrives as one total per platform. Work run by automated jobs belongs to no department, so nobody pays their share. |
| Is access under control? | HR knows who left. Unless data access is tied to HR, it can stay open after they go. |
| Where did access change? | Permissions pile up one at a time. Nobody reviews who gave what to whom, or whether it happened at night or at the weekend. |
| What data left? | A copy of customer records sent outside the company looks like ordinary work, unless someone checks. |
What a close look would find
- Some people who have left may still have access to your data. Unless someone checks each account against the leaving dates in HR, nobody knows to close it when they go.
- Some copies of production data may sit in test tables, labeled as ordinary data. Production’s access rules don’t follow the copy, so people outside the team that made it may read it.
- Some personal data may have been sent to storage your company doesn’t own. Each export is on record, but on each platform separately, where it’s easy to miss.
- Some of your data platform bill may belong to no team. The bills show cost by platform, not by the people and teams whose work ran it up.
Data Control Tower checks for all four every night. It ranks what it finds and names the team that should fix each one.
Leaders get answers they can defend. Data owners get a list of their own.
The leader who answers for the data
Usually a Chief Data and Analytics Officer, or the head of data governance.
- Governance they can show, beyond the catalog. Who used the data, what it cost and who can reach it, in one place.
- Each department sees its share of the bill. It is an estimate, based on the work its own people ran. Automated jobs count once you name the department they serve.
- Evidence for access reviews. Who can reach what, who changed it and when, and who left but still has access.
- A ranked list of what to do next. Each item says why it matters and which team acts.
The person who owns the data
Data owners and stewards. Governance happens in the data they look after, one decision at a time.
- They see who reads their data, and how many more people could.
- They see who can reach it today, not only the people they once approved.
- They learn if any of it left the platform, or was copied outside production.
- They get their own list of what to do, for only the data they look after. They choose what to watch.
Six questions, answered for the people who ask them
For leadership and department heads. Who uses the data, how often, and which tables nobody has read in 90 days.
UsageFor the data team. Which data each department reads, the tools people use, and the data they combine by hand.
Usage · Query patternsFor finance. The monthly bill per platform, and each department’s share of it.
CostFor security. Admin power, leavers who still hold access, refused attempts, and service accounts used by people.
Access · Service accountsFor platform admins. Who gained access, who lost it, who made each change, and whether it was out of hours.
AccessFor security and data owners. Exports and downloads of personal data, how many records, and whether they left the company.
Access · My assetsEach role starts on the page that answers its question. Home puts all six answers on one page for leadership, and Actions gives each team its own list.
Every answer ends in what to do next, and which team does it.
Six questions, one card each, with the change from last month and the top action for each.
People, departments, data areas and tables on every platform you connect, and the tables nobody reads.
What people do, the tools they use, when they work, and which data they combine by hand.
The monthly bill per platform, and each department’s share.
Open risks ranked, every access change with who made it, admin power, leavers, refusals, and data that left.
What each automated account does and costs, and, on Redshift and Snowflake, when a person uses one.
Work done outside production, and test tables built from sensitive production data.
Every action from every page, ranked and grouped by who acts. Each closes on its own when the next day’s data shows the fix.
The data owner’s view: every question, asked about only the tables they look after.
Every number carries a note saying how it was calculated, so anyone can see which records produced a finding before acting on it. Every page works at phone width.
In your Azure. Read-only. No platform passwords stored.
Data Control Tower deploys into your own Azure subscription from a template your IT team can read before it runs. Every day it reads the records your data platforms already keep, and it never writes to them. Nobody changes how they work.
- Amazon RedshiftQuery history, grants, exports
- AWS billCluster and storage cost
- DatabricksIts system tables
- SnowflakeIts usage history
HR file and catalog
Departments, leavers, tags, owners
Daily collectors
Read-only
Store
History kept past Redshift’s 7 days
Findings ranked
Data leaving and leavers first, then admin power and unusual grants
Query text can carry customer values, so it is read only to find exports, grants and refused attempts, then dropped. What is kept is who, when, which tables, how many rows, and a fingerprint of the query with no values in it. All of it stays in your Azure subscription, encrypted at rest.
| Your security and platform teams will ask | The answer |
|---|---|
| Where does it run? | In your own Azure subscription, deployed from a template your team can read before it runs. |
| What can it change? | Nothing. It reads and never writes. On AWS, one read-only role that your admin creates with our script. On Databricks, read access to the system tables, granted by your admin. On Snowflake, a read-only role and user, created by your admin from the script shown at setup. |
| Are platform passwords stored? | No. Redshift and Snowflake are read with the product’s own Azure identity. On AWS, a read-only role trusts it, so no AWS key is stored and Redshift stays private. Snowflake trusts it directly, so no Snowflake password or key is stored. No Databricks token is stored either. The only stored secret is your catalog’s token, in your Key Vault. |
| What is never kept? | Raw query text. It is read to find exports, grants and refusals, then dropped. |
| How much history on day one? | A full year from Databricks. Seven days from Redshift, unless your Redshift audit logging to S3 is on. Redshift’s 90-day measures fill in over the first 90 days, and the page marks them until then. On Snowflake, the past 90 days load a little each night, over about the first week. Access changes are tracked from the day it is installed on Redshift and Databricks, and from up to 90 days before on Snowflake. |
| How are people matched across platforms? | By email, against your HR file, so a person who uses more than one platform is one person, not several logins. Logins that don’t match stay visible as unmatched, never dropped. |
| Is cost per person exact? | On Redshift and Databricks, it is an estimate, and the page says so: each hour of a shared warehouse is split by each account’s share of query time. On Snowflake, each query’s cost comes from Snowflake’s own records. Service accounts go to the department they work for, and idle time to its own line. |
A fixed price, with four reviews a year included
Installed, set up, reviewed
- For any two of Amazon Redshift, Databricks and Snowflake
- Installed in your own Azure subscription
- Setup: your HR file, catalog tags, service-account owners, the storage you own and your working hours
- All nine pages
- Four quarterly reviews, each ending in a ranked list of actions
Renewal
- Continued use of Data Control Tower
- Updates as Redshift, Databricks and Snowflake change
- Four quarterly reviews, each ending in a ranked list of actions
Running one platform, or all three? Ask us for the price. No software license fee. You pay your own cloud bills for the compute it uses: Azure, and a few minutes on each data platform each night.
Available on Microsoft Marketplace. Listed after passing Microsoft’s certification review, so your IT team installs it from the Marketplace like any other Azure application.
Applications close November 20
0days00hours00min00sec
Put Data Control Tower to work on your own data
We’re offering a free pilot of Data Control Tower to one company. Data Meaning charges no fee, and there is no license fee. You pay only your own costs, on your own bills: the Azure resources it runs on, and a few minutes of compute on each data platform it reads each night. We install it in your own Azure subscription and set it up. It only reads, and it stores no platform passwords.
What you need, and when it is not for you
You need
- One, two or all three of Amazon Redshift, Databricks and Snowflake, and an Azure subscription to install into
- Read-only access to each platform’s system records, granted by your platform admins after install. No passwords
- For Snowflake, your account identifier and price per credit
- Your HR file: department, manager, employment type, and start and end dates
- Your catalog tags: which tables hold personal data, and which areas are production
- Who owns each service account, and the servers it should run from
- The storage buckets you own, so an export anywhere else reads as outside
- Your working hours and time zone, so out-of-hours changes are marked correctly
It is not for you if
- You run Snowflake Standard Edition and need to see which tables people read. That needs Snowflake Enterprise Edition
- Your data is on none of Amazon Redshift, Databricks or Snowflake today
- You want pipeline health or data quality monitoring
- You want a tool that blocks exports or masks data. It shows and ranks; it does not block
- You want it to change who has access. It reads and never writes. For access requests and approvals, see Data Access Concierge.
Pick a question. See Data Control Tower answer it.
Each short demo shows the result first, then how Data Control Tower found it, on a fictional company’s data.
Start with the whole picture
One view across your data platforms
One page answers all six questions across your data platforms, from whether the data is used to what it costs and who can reach it. Each answer names the next step and the team that takes it.
One ranked to-do list
Every finding, from leavers to test copies to cost no team pays, becomes one list ranked by risk, split by the team that acts, with why each item matters and the steps to take.
A data owner’s view
One page for the tables an owner looks after: who read them, who can reach them today, who exported data from them, and which test tables hold copies.
Who uses your data, and what it costs
Who uses the data, and how
Which people and departments read the data, which tables nobody has read in 90 days, the programs people query from, and the tables people keep combining by hand.
Cost per person and per team
An estimate of what each person’s queries cost, added up by department so the cost can be charged back. Service accounts with no department named to pay for them are flagged.
Who can reach it, and the proof for your auditor
Leavers who still have access
Every account is checked against the leaving dates in your HR records. See which leavers still have access, whose account was used after they left, and what to do today.
Risky access changes
Every access change, with who made it and who got it. Flagged for a look: access given to a contractor, access people give themselves, changes outside working hours and new admin power.
Service accounts used by people
When a person signs in to a service account, they get all its access, under its name. On Redshift and Snowflake, see the days that happened and the tables read that its job never touches, plus access given to service accounts outside working hours.
Governance Evidence Pack
A separate module, in early access: SOX or HIPAA audit evidence built from your platforms’ own records. Someone other than its author approves it, and the auditor gets the queries to re-run every finding.
Where copies of your data end up
Production data copied into test and development
Every test and sandbox table built from production data, what each copy holds and who reads it. Flagged: personal data labeled as ordinary, or read by people outside the team that built it.
Data that left the company
Everyone who exported or downloaded personal data, and how many records. Flagged: repeated downloads, and exports to storage your company doesn’t own, with the leaving date if the person has left.
What leaders ask us first
We already have a data catalog.
Our data platforms already show some of this.
Our security team already has a tool for this.
Can we trust a number about a person?
Our IT won’t let an outside firm install something in our cloud.
Why quarterly reviews? Just give us the software.
Can it handle access requests too?
Know what happens to your data: who uses it, what it costs, and who can still reach it.
The intro call is a demo of Data Control Tower on a fictional company’s data. You’ll see those answers on one page, then the ranked list of what to fix and which team acts on each item. You’ll also see the page a data owner gets for the tables they look after. You bring nothing. To see it on your own data, apply for the free pilot.
Data Meaning is an independent consultancy. Amazon Redshift, AWS, Databricks, Snowflake and Azure are trademarks of their respective owners.