Where the Build Money Goes

It is about sixty days of work

Not a website. A scheduling system that takes messages in, routes work to the right person, tracks whether they answered, and keeps the money straight. Every line below is something that has to exist before the thing is usable on a Tuesday morning.

What gets built	Why it exists	Days
Accounts and permissions	You, a coordinator later, different access	3
Database and data model	Members, properties, jobs, vendors, fees	3
Inbound texts become records	A member's text turns into a tracked request, threaded to them	5
Dispatch board	Everything open, with the response clock running	5
Contractor bench	Vendors, documents, and automatic suspension when insurance lapses	4
Offer engine	Send an offer by text, parse the reply, expire it, re-route to the next vendor	6
Quote review	Log it, compare against past jobs, approve or send back	4
Scheduling and close-out	Book a window, record what was actually invoiced	3
Fee ledger	Accrue 10%, invoice it, chase it, reconcile it	4
Member and property records	The home file — roof age, HVAC age, repair history	3
Maintenance reminders	Nightly job that finds what is coming due and texts about it	3
Photos in and out	Picture messages, stored against the job	2
Stripe sync	Who is actually a paying member right now	2
Carrier registration	Business texting has to be registered or it gets silently filtered	2
Testing and fixing	Roughly a fifth of any build that you intend to trust	8
Deploy, monitoring, handover	Live, watched, and documented enough that someone else can take it over	3
Total	One developer, about twelve weeks	60

The price is about who builds it, not what gets built

Sixty days is sixty days. What moves the number by a factor of three is the hourly rate of the person doing them. This is the single most useful thing to understand before you take quotes from anybody.

Who builds it	Rate	This scope	What you are trading
Offshore agency	$50–70	$24k–34k	Cheapest. Timezone gap, and quality varies enormously between shops
Senior contractor, AI-assisted	$70–110	$34k–53k	One accountable person who can move fast. Bus factor of one
US agency	$125–175	$60k–84k	Project management, continuity, someone to sue. You pay for all three

A correction worth making.

The proposal originally said $25k–40k for this. Working it out line by line, that was low — it only holds at the cheapest end of the offshore range. The honest figure for a senior contractor is $35k–50k, and I would rather tell you that now than discover it in month three.

Why it costs more than a website

A website shows information. This one moves work and money. A brochure page that renders wrong is ugly. A dispatch board that loses a request means a member sat unanswered on the day their water heater failed.
Every edge case is a real person waiting. The contractor who never replies. The text that says "sure thing" instead of Y. Two vendors accepting the same job four seconds apart. Each of those is a decision someone has to code, and the pleasant version of the software is mostly made of them.
The money has to reconcile. A fee that accrues on the quoted price rather than what was collected, or a Stripe cancellation the system never noticed, is an invoice argument with a vendor you cannot afford to lose. Getting that right is careful, unglamorous work.
Compliance is not optional and not free. Registered business texting, consent records, insurance certificates that suspend a vendor automatically when they expire. None of it is visible to a customer and all of it is the reason the business does not blow up.
Testing is a fifth of the build. It is the fifth that decides whether you trust the fee number at the end of the month without checking it by hand — and if you do not trust it, you will check it by hand, and you have bought nothing.

Three ways to pay less

All of them are real options. I would take any of them over spending $40,000 before you know the business works.

Buy only the offer engine

$12k–19k

Inbound texts become tracked requests, offers go out by text, replies get parsed, unanswered offers expire and re-route to the next vendor on the bench. A minimal board to watch it happen. Nothing else.

You get: The piece that actually gives you your hours back — no more calling three plumbers to find one who is free
You keep: Doing quotes, scheduling, fees and the home file the way you already do
Why first: It is a fifth of the scope and most of the time saved. Build the rest when it still hurts

No-code first

$3k–8k

Airtable for the data, Softr or Glide for the board, Make or Zapier to wire the automation, Twilio for the texting. A working system in two or three weeks.

You get: Almost all of the function, almost immediately, for a tenth of the money
You give up: Speed past roughly a hundred jobs a month, and the ability to export the logic when you outgrow it
Watch for: Tool subscriptions of $150–300 a month, and a rebuild later if this becomes the real system

Stay off the shelf longer

Jobber for dispatch, OpenPhone for texting, a spreadsheet for the bench. What you are launching on anyway.

You get: Zero build cost and zero risk while you find out whether members renew
You give up: Nothing, until roughly 40–60 jobs a month, when the seams start to show

What I would actually do

Launch on off-the-shelf tools and stay there until tracking jobs by hand genuinely costs you hours. Probably six months. Possibly a year.
Then buy the offer engine alone, at $12k–19k, because dispatch is where the hours go. Fund it out of membership revenue, not out of savings.
Then build the rest of the console only if it still hurts. By then you will know which three screens you actually wanted, and the $35k–50k version may turn out to be a $20k version.
The member portal comes last, at $12k–20k, and only once members are asking to see their own history — which they will not do in year one.

The number that matters more than any of these.

At 350 members this business does roughly $116,000 a year. Spending $40,000 of it on software before you have the members is not a technology decision, it is a solvency decision. Every option on this page exists so you do not have to make it.