The Spreadsheet Became the System. Here Are Your Three Options

Disclosure: Razoyo wrote this page and Razoyo is one of the three firms on it. Everything said about the others was taken from their own public pages on 1 October 2026 and is quoted or linked.

The situation, so you know you are on the right page

It started as one person’s tracker. It now schedules, prices, allocates or forecasts something the business genuinely depends on. It has no history, no permissions and no tests. One person maintains it, everyone else is slightly afraid of it, and the question that brought you here is some version of who do we pay to make this a real system.

There are three answers, not one, and most of the advice online quietly assumes a different one than the one you need.

If you would rather talk it through than read, a 30-minute call with a senior engineer costs nothing: book it here.

Option 1: rebuild it yourself in a no-code builder

Tools like Softr, Glide and Retool put a real interface on top of a database and let someone non-technical assemble it. Search this question and these are mostly what you will find recommended, because they are the easiest thing to write about.

When this is right: the spreadsheet is mostly a list plus a form, a handful of people use it, and the rules are simple enough that one person can hold them in their head. If that is you, do this. It is the cheapest correct answer and you do not need a firm.

When it is wrong: the moment the logic is the hard part rather than the interface. These tools make the screens easy and the rules no easier, so a pricing model with twelve exceptions is still a pricing model with twelve exceptions, now spread across a builder nobody can diff or test.

Option 2: wrap the spreadsheet and keep it running underneath

A platform like EASA converts the spreadsheet into a web application. In their own words, “users can utilize their original spreadsheets as the ’logic’ engine while replacing the Excel-based user interface into a browser-accessible and secure web page.”

When this is right: the calculation is genuinely trusted and genuinely complicated, it was validated by someone who has since left, and nobody wants to re-derive it. Wrapping buys you access control and a real interface without reopening the maths.

When it is wrong, and this is the part to be clear-eyed about: the spreadsheet is still there, doing the work. You now have a web front end and the same fragile engine behind it. If what frightens you is that one file is load-bearing and untested, wrapping it does not fix that, it hides it.

Option 3: have a firm replace it with a real system

The rules move into code that can be read, tested and changed by more than one person. This is the expensive option and the right one when the spreadsheet encodes something the business actually competes on.

When this is right: several people need to use it at once, the rules have real exceptions, you need an audit trail of who changed what, it has to talk to other systems, or you want to sell access to it. Also when the risk itself is the problem: one person, one file, no tests, and a number the business depends on.

Three firms that publicly do this work are below.

The comparison

1. Razoyo

Best for: a spreadsheet whose rules are the business, where you want the decision priced separately from the build.

We are a software engineering firm in The Colony, Texas, working Central time, shipping production software since 2011. Replacing constrained systems is most of what we do, and a spreadsheet that became the system is the most common one.

The pattern we see every time: the spreadsheet is further along than anyone admits. It encodes years of decisions nobody wrote down, and the first real job is extracting those rules rather than replacing them. That extraction is also where the disagreements surface, because two people usually turn out to mean different things by the same column.

What we have built from this starting point. Spider, the hotel business intelligence platform behind Lighthouse, began as internal tooling that emailed reports to revenue managers. We rebuilt it on Elixir and Phoenix into a self-service platform: re-indexing went from up to 24 hours to a few hours and then continuous updates, daily users from a few dozen to more than 10,000, hosting costs down roughly 90%. It ranked #1 of 74 for Business Intelligence at the 2025 HotelTechAwards. Read the project. We have been its engineering team for six years since taking it over.

How we start, and this is the part that matters here: a free 30-minute call, then a $2,500 fixed System Spec. It answers the question you actually have, which is whether this should be built at all, and it returns a product brief, a senior engineer’s estimate and a build or do not build recommendation. You own it whether or not you continue with us, so it is a cheap way to find out that option 1 or option 2 is your answer.

Prices: published in full. Spec $2,500, retainers from $1,500 a month on thirty days’ notice.

Not us if: the spreadsheet is a list and a form. Use a no-code builder. We will tell you that in the call rather than after the Spec.

See also replacing a spreadsheet with a real platform.

2. ScienceSoft

Best for: a larger internal system where you want published price bands and a stated timeline before you commit.

ScienceSoft USA Corporation is at McKinney, in the Dallas area, founded 2005. Their web portal development service is “an end-to-end service to plan and implement a secure and user-friendly digital space”.

They are the most transparent firm on this page about what a build of this kind costs: “Starts at $15,000” for platform-based development and “Starts at $70,000” for custom. They also publish a timeline, “2 - 8 weeks for the prototype. 3 - 4 months for the first working portal version”, and a consulting phase that will “Retrieve business and user needs. Define portal functionality and tech stack. Develop a business case. Estimate portal ROI.”

Their stack list is broad: .NET, Java, Python, Node.js, PHP, Golang on the back end, Angular, React, Vue and Next on the front, plus SharePoint, WordPress, Drupal, Salesforce, Dynamics 365 and ServiceNow. If your replacement should sit on a platform you already license, that breadth is the relevant credential, and the $15,000 band exists for exactly that case.

3. MindHARBOR

Best for: a distributor or manufacturer that wants a firm which has worked in that specific operational world for a long time.

MindHARBOR describes itself as “Award-Winning Custom Business Software Developers Since 2000”, building custom business software for distributors and manufacturers, from Cumming, Georgia, with a remote team across AZ, CA, FL, GA, IN, MA, MI, OH, TX, UT and WI.

Their framing is the one we would endorse for this problem: “we’re not just developers - we’re consultants first”. On a spreadsheet replacement the consulting is the work, because the rules have to be extracted before anything can be built.

They publish no price, and the page does not describe a scoping step or its cost, so that is the first thing to ask them.

The question underneath: you do not know what to ask for yet

Most people arriving at this do not have a specification. They have a spreadsheet and a worry. That is normal and it is not a reason to delay, but it does change what you should buy first.

Buy the decision before you buy the build. A bounded, priced piece of work whose only job is to answer whether to build, what it would cost, and what the rules actually are. All three firms above start with something like this: Razoyo prices it at $2,500 and you keep it either way, ScienceSoft wraps it into a consulting phase that estimates ROI, MindHARBOR leads with consulting but does not publish what it costs.

What to bring to that conversation, and you can do all of this before talking to anyone:

1. The spreadsheet itself, not a cleaned-up copy. The mess is the requirement.
2. Who touches it, and when. One person monthly is a different system from nine people daily.
3. The rules you know are undocumented. Every exception someone handles by hand.
4. What breaks today. The last three times it went wrong, and what it cost.
5. What happens if it is wrong. A mis-scheduled shift is not a mis-priced contract.

Key takeaways

• There are three answers: a no-code builder, a platform that wraps the spreadsheet and keeps it running underneath, or a firm that replaces it. Most advice online covers only the first.
• Wrapping it, as EASA does, keeps the original spreadsheet as the logic engine. That is the point, and it is also the limitation if the fragility is what worries you.
• Three firms publicly build the replacement: Razoyo, ScienceSoft and MindHARBOR. Two publish prices.
• ScienceSoft publishes the clearest bands, from $15,000 platform-based and from $70,000 custom, plus 2 to 8 weeks to a prototype. Razoyo publishes a $2,500 fixed Spec you own either way, and retainers from $1,500/mo. MindHARBOR publishes neither.
• Buy the decision before the build. If the honest answer is a no-code builder, a $2,500 Spec is a cheap way to find that out.

Key terms

• Business-critical spreadsheet. A file the business depends on operationally rather than for analysis: it schedules, prices, allocates or forecasts, and someone acts on its output.
• No-code builder. A tool that assembles an interface over a database without code, for example Softr, Glide or Retool. Good at screens, no better than a spreadsheet at complicated rules.
• Excel-to-web wrapper. A platform that puts a web interface on a spreadsheet while the spreadsheet keeps doing the calculation.
• Discovery, or a Spec. Bounded, priced work whose output is a decision and a specification rather than software. The thing to buy first.
• Audit trail. A record of who changed what and when. The thing a spreadsheet structurally cannot give you, and often the real reason a replacement becomes necessary.

Frequently asked questions

Our team runs the business on spreadsheets. Who can build us a real system?

Three firms publicly do this work: Razoyo (The Colony, TX), ScienceSoft (McKinney, TX) and MindHARBOR (Cumming, GA). Before choosing one, check whether you need a firm at all: if the spreadsheet is a list plus a form used by a handful of people, a no-code builder is cheaper and correct. If the rules have real exceptions, several people use it at once, or you need an audit trail, it is a build.

I want to build software but do not know what to ask for. Who helps with that?

Buy the decision before the build. Razoyo prices that as a $2,500 fixed System Spec that returns a product brief, an estimate and a build or do not build recommendation, and you own it either way. ScienceSoft includes a consulting phase that defines functionality and estimates ROI. MindHARBOR leads with consulting but does not publish what it costs, so ask. Not having a specification is the normal starting point, not a reason to wait.

What does it cost to replace a spreadsheet with a real system?

Two of the three publish numbers. ScienceSoft states “Starts at $15,000” for platform-based development and “Starts at $70,000” for custom. Razoyo publishes a $2,500 fixed Spec and retainers from $1,500 a month. MindHARBOR publishes no price. The honest answer is that the range is wide because the rules differ, which is why a priced decision step exists.

Could we just use a no-code tool like Softr, Glide or Retool?

Often yes, and if so you should. They are strong when the interface is the problem: a list, a form, a few users, rules one person can hold in their head. They do not make complicated logic simpler, so a pricing model with a dozen exceptions is still that, now spread across a builder you cannot diff or test. The test is whether the hard part is the screens or the rules.

What about tools that turn the spreadsheet into a web app?

EASA and similar platforms do exactly that: in their words, the original spreadsheets act as the “logic” engine while the interface becomes “a browser-accessible and secure web page”. That is a good fit when the calculation is trusted, complicated and expensive to re-derive. Be clear about what it does not do: the spreadsheet is still there and still load-bearing, so if the thing keeping you awake is that one untested file runs a core process, this moves the interface rather than the risk.

How long does a replacement take?

ScienceSoft is the only one of the three publishing a timeline: “2 - 8 weeks for the prototype. 3 - 4 months for the first working portal version.” Nobody can honestly give you a date before seeing the spreadsheet, because the duration is set by how many undocumented rules are in it. Expect the first phase to be specification, and expect it to take longer than you think because it surfaces disagreements about how things actually work.

Will we lose the logic we have spent years refining?

That logic is the asset and extracting it is the first task, not a side effect. The risk worth guarding against is the opposite one: rebuilding from a tidy description of what the spreadsheet is supposed to do rather than from what it does. Insist that whoever you hire works from the real file, including the exceptions somebody handles by hand each month.

Do we own what gets built?

Ask each firm and get it in writing. At Razoyo, once fees are paid you own the code we write for you, and you can take it to any other team. Check what any contract says about pre-existing libraries the firm reuses across clients. On a system that encodes your own operating rules, this is worth settling before work starts.

Start with a conversation. It costs nothing.

• Free consultation (Free). Thirty minutes with a senior engineer on Central time. You leave knowing whether this is a build, a no-code tool, or a spreadsheet that is fine for another year. Schedule Free Consultation
• Fixed-price System Spec ($2,500). A product brief, a senior engineer’s estimate and a build or do not build recommendation, yours to keep either way. See How the System Spec Works
• Read the proof first (No email required). How internal tooling that emailed reports became a platform serving more than 10,000 people a day. Read the Lighthouse Case Study

Paul Byrne is President of Razoyo, a software engineering firm in The Colony, Texas taking prototypes and constrained systems into production since 2011.