Building Internal Tools Faster Than Spreadsheets
When sheets break, operations break. How to ship a light internal tool with auth, roles, and workflow that deploys faster than living in Excel chaos.
Ali Mortazavi
Founder, Paradise Code
How far is Excel still your friend?
For ten rows and one owner, a sheet is perfect. When three people edit at once, “final2_fixed” versions appear, and formulas change without tests, human-error cost has already exceeded software cost.
Exit signals: weekly inventory mismatches, no change history, all-or-nothing file access, and reports only one person can build because “they know the sheet.”
What an internal-tool MVP means
Phase one is not replacing the company. Pick one painful flow—B2B order intake, leave approval, supplier ticket tracking—and ship form, list, and statuses.
Good scope: one primary role, 5–7 key fields, three statuses, and a simple notification. If phase one needs more than 4–6 weeks, you’re probably building an ERP, not a tool.
Data, access, source of truth
A small database with backups beats five unsynced sheets. Add roles on day one even if simple: viewer, editor, approver.
CSV import is mandatory so migration isn’t traumatic. Unvalidated free-for-all imports just move Excel chaos into a database—enforce minimum rules.
Why “faster than sheets” is possible
Because you don’t start with heavy custom visual design: tables, forms, filters, and activity logs are repeated patterns. With Next.js + NestJS + a UI kit, the skeleton stands in days.
Real speed comes from scope cuts, not framework magic. Park every “one more report” in phase-two backlog unless it blocks daily work.
Org change, not only software
If managers still demand a parallel sheet export as the real system, the tool dies. Set a cutover date, make the old sheet read-only, and appoint an ops champion to train peers.
Success metrics: task time, rework/error count, and fewer “send me the file” requests. If those don’t move, don’t restyle the UI—revisit the flow.
A practical path for Iranian teams
Week 1: discovery and flow map. Week 2: data model and access. Weeks 3–4: main-flow UI + import. Week 5: pilot on real data. Week 6: friction fixes and cutover.
Paradise Code builds internal tools like real products—with RBAC and audit logs—because ops data is often more sensitive than a marketing About page.
Frequently asked questions
Airtable/Notion instead of custom?
Fine for tiny teams and low-sensitivity data. Precise access control, accounting integration, or high volume usually wants custom or at least your own backend.
Do we need mobile?
Yes for field users on 2–3 key tasks; otherwise desktop-first is enough.
How do we avoid an endless project?
Declare one flow Done, write acceptance criteria, and priority-score every new request.
Insights
Need these ideas implemented in your product?
Paradise Code supports you from consult to full delivery.