The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Authenticated does not mean isolated.
I mocked up a small multi-tenant operations SaaS today: three companies, four roles, and 60 synthetic records. Every route required a valid session. Every role check passed. One export endpoint still returned records belonging to all three companies.
The tenant filter was being applied after aggregation.
More authentication middleware would not have fixed it. The solution was to move tenant identity into the data-access contract:
• tenant_id required by every repository method
• database row policies as a second boundary
• composite keys that include the tenant
• integration tests that deliberately query as the wrong tenant
• background exports carrying tenant context from creation through storage
The uncomfortable part was that the UI looked completely correct. Only the exported CSV exposed the failure.
Many SaaS teams test, “Can this user access the endpoint?” Far fewer test, “Can this valid user ever see another customer’s data?”
Which layer do you trust for tenant isolation: application, database, or both?
#SaaSSecurity #BackendEngineering #MultiTenant
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
End-to-end retail analytics pipeline that turns raw, messy CSV exports into business decisions, built in Python and SQL and runnable with one command.
What it does:
Cleans and validates 6 relational tables, fixes 585 data-quality issues and logs every action
Loads a SQLite warehouse with keys, constraints and window-function SQL
Customer analytics: RFM segmentation, cohort retention, CAC / LTV by channel, product Pareto, market-basket analysis
Churn prediction model validated on future data (ROC-AUC 0.71, no data leakage)
6-month revenue forecast, best of 3 models by backtest (4.2% MAPE)
Auto-generated Excel report (live formulas, native charts) and an interactive Streamlit dashboard
18 automated tests
Tallisto: brand identity for an AI accounting SaaS that closes the monthly books. A concept project.
The mark comes from a medieval tally stick. In 12th-century England a debt was recorded by cutting notches into a stick, then splitting it in two. Each side kept a half, and if the halves matched, the books agreed.
That is what Tallisto's agent does. It matches every bank line to the ledger, and a person signs it off. So the logo is two sticks, the stock and the foil, carrying the same two cuts.
One rule holds the system together: Lamp Green only appears where a person has approved the agent's work. It's never used on buttons or as decoration.
The rest is calm on purpose. Ink and Paper as the base, Instrument Sans for words, JetBrains Mono for figures, and a "ledger of tallies" pattern that never sits behind text. You can see it here on the website, the close dashboard, a mobile sign-off screen, business cards, an office wall and an Underground poster.
Would a finance team trust a tool more if its logo had a reason behind it, or do they only care about the numbers?
I design brand identities and build them into Webflow sites. Open to SaaS and fintech projects.
The tally stick story is the kind of thing a founder retells in every investor meeting, that's a brand doing sales work. "Lamp Green only where a person approved" is my favourite rule here, and it would carry beautifully into a board report too: green only on the numbers that are signed off.
At handover, check that the source files and assets actually landed on your side. And get short notes on where everything lives.