Replace a spreadsheet when several people edit the same data, need to see different parts of it, and copy it between files and emails. Before you build, try the cheaper rungs: a better spreadsheet, Airtable, a low-code tool or an off-the-shelf CRM. Build a custom internal tool when your process involves outside parties or rules those tools make you fake.
I build internal systems at DodaCraft, and I still send some people back to a well-organised spreadsheet. This post is about telling the two situations apart.
Short answer: when should you build an internal tool?
Build when the workflow is specific to your business, several parties touch the same records, and per-seat SaaS fees over three years approach the cost of a build. If your process looks like everyone else's sales pipeline, buy a CRM and spend the money elsewhere.
What are the signs your spreadsheet has outgrown itself?
- Two people edit the same file and overwrite each other, or keep "final_v3" copies.
- You hide columns or make separate copies because some people must not see some data.
- Data arrives by email or PDF and someone types it into the sheet by hand.
- Nobody can tell who changed a value, or when.
- Status lives in someone's head: "Ask Tom, he knows which ones are waiting for the bank."
- Clients or partners send you the same information several times because there is no form.
One or two of these is a reason to tidy the spreadsheet. Four or more, and the spreadsheet is costing you staff hours every week.
What are the options between a spreadsheet and custom software?
Think of it as a ladder. Climb one rung at a time and stop where the problem goes away.
| Option | Cost model | Setup effort | Where it breaks |
|---|---|---|---|
| Better spreadsheet (validation, protected ranges, one shared file) | Usually included in tools you already pay for | Hours | Permissions per row, outside users, audit trail |
| Airtable or Notion | Per editor per month | Days | Complex permissions, external parties, cost as the team grows |
| Low-code (Retool, Softr) | Per builder and per user per month | Days to weeks, someone must build it | Unusual logic, document processing, vendor lock-in |
| Off-the-shelf CRM (HubSpot, Pipedrive) | Per seat per month, sometimes an onboarding fee | Weeks of configuration | Processes that are not a sales pipeline |
| Custom internal tool | One-off build, hosting, maintenance | Weeks to months | Small teams with standard processes, where it is overkill |
Some prices from the vendors' pages at the time of writing. Airtable's pricing page lists Team at $20 and Business at $45 per user per month, billed annually, counted for users with edit access. Retool's pricing page lists the Business plan at €46 per builder and €14 per internal user per month on annual billing. HubSpot's Sales Hub page shows Professional starting at $90 per seat per month, plus a required one-time onboarding fee of $1,500.
What did Rentclub need that a CRM could not give?
Rentclub is a car leasing platform that connects car dealers, clients and banks. Before the build there was no single place for applications. Clients filled forms by hand, data travelled through several systems, and answers were slow.
A sales CRM models one company selling to customers. Rentclub has three outside parties with different jobs. We built it from scratch with Next.js, Strapi and PostgreSQL:
- Dealers upload cars.
- Clients apply for a specific car or a specific amount.
- A separate module processes bank statements and passes the data to banks.
- Banks submit offers.
- Staff manage applications in an internal system.
Analytics shows where clients drop their application, so the team knows which step to fix. The lesson I took from it: when data travels between several parties, standardise it at the very start. Agree on field names, formats and statuses before the first screen is designed. Changing the data shape later means changing it for dealers, clients, banks and staff at once.
How do you scope the first version?
The first version should do one workflow end to end and nothing else. Write these down before you talk to any developer:
- Who sees what. List every type of user and, for each, which records they can see and change. Outside parties usually see only their own records.
- One workflow. Pick the one that costs the most hours today, such as "application arrives, gets checked, goes to approval".
- The data model. The things you track (clients, applications, documents) and how they connect. Standardise formats now.
- Permissions. Roles, not individual exceptions. Exceptions multiply.
- Audit trail. Who changed what and when. Cheap to add at the start, painful to retrofit.
- What stays manual. Name the steps you will not automate yet. That keeps the scope honest.
If part of the workflow is reading documents or emails and typing them into fields, that step may suit an AI step inside the tool. My AI automation work covers that piece.
How do you compare SaaS and custom costs over 3 years?
Use the same horizon for both. Three years is long enough to include maintenance and short enough to be realistic about your process staying the same.
SaaS: per-seat price × team size × 36, plus onboarding fees and paid add-ons.
Custom: build cost + hosting × 36 + yearly maintenance × 3. Ask for maintenance as a written figure in the quote, not "as needed".
A worked example with list prices. Eight editors on Airtable Business: 8 × $45 × 36 = $12,960. Five seats on HubSpot Sales Hub Professional: 5 × $90 × 36 = $16,200, plus $1,500 onboarding, so $17,700. Retool Business with two builders and ten internal users: (2 × €46 + 10 × €14) × 36 = €8,352, and someone still has to build the app.
The custom side has no list price, so put your written quote into the formula. Then look past the totals. If the SaaS tool matches your process, it wins, because a company that is not you maintains it. If you would spend three years working around it, the custom tool wins.
Also count what does not show on an invoice: hours spent copying data, and the extra seats you will need as the team grows.
How much does internal tool development cost?
The range is wide, because "internal tool" covers both a single admin screen for one team and a platform like Rentclub, where dealers, clients, banks and staff work in the same system with document processing in between. A first version with one workflow and a few roles is a small project. Every extra outside party, integration and document type adds to it.
At DodaCraft, internal tools start from €999 excluding VAT for a scoped tool, with the quote and timeline in writing before work starts, and you own the code and hosting accounts. The first 30-minute call is free, and if a spreadsheet with proper validation would solve it, that is what I will tell you.
If your bottleneck is client bookings rather than internal data, read whether to build or buy a booking system.
FAQ
When should I replace a spreadsheet with an app?
When several people edit the same data, when you need different people to see different rows, or when mistakes come from copying data between files and emails. One person tracking one list is fine in a spreadsheet.
Is a custom CRM better than HubSpot or Pipedrive?
Not for a standard sales pipeline. Off-the-shelf CRMs are cheaper to start and well supported. Custom wins when your process involves outside parties, unusual data or steps that a CRM forces you to fake with custom fields.
How much does internal tool development cost?
A scoped first version with one workflow starts from €999 at DodaCraft, excluding VAT. Systems with several user types, integrations and document processing cost considerably more.
How do I compare SaaS and custom costs over three years?
Multiply the per-seat price by your team size and by 36 months, then add onboarding fees. For custom, add the build cost, hosting for 36 months and yearly maintenance. Compare the two totals and the fit.
What should the first version of an internal tool include?
One workflow, done end to end, with the data model, roles and permissions, and a log of who changed what. Everything else waits until people use it daily.
Tick what you need and see the starting price straight away.
Founder of DodaCraft, developer. I build websites, online stores and internal systems for businesses in Lithuania and abroad. I taught teenagers to programme for 7 years.
