Turning a business-critical Excel tool into a web app: cost and steps

The verdict in three sentences

A 28-tab Excel workbook that produces every quote in an engineering consultancy and that only one person can maintain is an operational risk, not an office tool. Turning it into a web application with migrated formulas, user permissions, version history and PDF export costs, as a 2026 order of magnitude, GBP 10,000 to 30,000 with a UK agency, over 6 to 12 weeks. The decision comes down to one number: what a formula error in a client proposal costs you.

The real cost of a spreadsheet that became critical

In a 35-person engineering consultancy in Bristol (structures, building services, energy), the pricing workbook has grown for ten years. It holds rate schedules, coefficients by building type, hourly rates by role, VBA macros and hidden tabs. One senior engineer is its only keeper. When he is away, proposals wait or go out on an old version.

Spreadsheet research, notably by Raymond Panko, estimates that about 88 % of complex spreadsheets contain at least one error. It is not a skills problem: it comes from copied cells, ranges not extended and links to renamed files.

Moving to a web app is not about looking modern. It is about making the calculation reliable, shared and traceable.

The steps of an Excel to web app project

The project runs in five stages. The first one is the most underestimated: understanding what the workbook actually calculates.

• Workbook audit (1 to 2 weeks): mapping the 28 tabs, cell dependencies, macros and reference data. Dead tabs and contradictory formulas almost always turn up.
• Data model and calculation rules (1 to 2 weeks): rate schedules become tables, coefficients become admin settings, formulas become tested rules.
• Build (3 to 6 weeks): guided quote entry, server-side calculation, role-based permissions, version history for each proposal.
• Parallel testing (1 to 2 weeks): 20 to 30 real jobs are priced in both Excel and the app and compared to the penny.
• Go-live and training: half a day per team, then the workbook becomes read-only.

The low end fits a clean workbook with few macros. The high end covers files with heavy VBA, ERP imports or several approval chains.

Options to weigh before building

Not every firm needs custom software. A well-structured, locked Microsoft 365 workbook can serve a five-person team. No-code tools fit simple processes. A custom build is justified when the calculation is your know-how and several people need it at once.

The same logic applies to contract registers kept in a spreadsheet: see our piece on a contract renewal alerts and e-signature tool.

Mini case study

James, managing director of a 35-person structural engineering consultancy in Bristol, sends 200 proposals a year with average fees of GBP 15,000. In 2025, two proposals went out with an outdated coefficient, underpriced by 6 %, so GBP 1,800 of margin lost across those two jobs.

His business case:

• Errors avoided: 2 proposals a year at GBP 900 of lost margin each, so GBP 1,800.
• Proposal formatting: 45 minutes cut to 10, so 35 minutes saved on 200 proposals, about 117 hours at GBP 50 loaded, so GBP 5,850.
• Proposals unblocked during absences: 3 weeks a year, about 12 proposals delayed, 2 of them lost for missing the deadline, so GBP 30,000 of fees at stake.
• Project: GBP 20,000, maintenance GBP 2,500 per year.

Without counting lost bids, the direct gain is GBP 5,150 net of maintenance per year, a payback of about 3.9 years. If the app saves a single GBP 15,000 job a year, payback drops to about 12 months.

FAQ

Can we keep Excel running in parallel during the transition?

Yes, and it is recommended for 4 to 6 weeks. The workbook is then set to read-only and archived, so that two sources of truth never coexist.

What happens to our VBA macros?

They are rewritten as server-side calculation rules with automated tests. A workbook with 1,000 lines of VBA typically adds 2 to 4 days of audit and GBP 3,500 to 7,000 to the project.

Can users change the coefficients themselves?

Yes, through an admin screen limited to 2 or 3 people. Every change is time-stamped and attributed, which fixes the missing history problem in Excel.

Does the app work on site?

The web interface works on a tablet over 4G. For full offline use, plan a dedicated work package of 2 to 3 extra weeks.

How long until we are operational?

Between 6 and 12 weeks depending on workbook complexity. Parallel pricing on 20 to 30 real jobs is the step that earns the team's trust.

Let's scope your project. Send us your workbook (anonymised is fine) and your number of users, and we will return a dependency map, a work-package breakdown and a 6 to 12 week plan. Detailed quote within 48 h. WhatsApp +221 77 596 93 33.