Over three years, a stacked proptech budget usually costs a small firm more than a single custom automation project would, and a single custom automation project almost never replaces the whole stack. Both sentences are true at once, which is why the three-year math trips up so many principals: they compare one subscription to one build, when the real decision is between a growing stack of subscriptions and a fixed build that carries a maintenance tail. This piece gives you the model to run it for your own firm, the point where the two cost lines cross, and an honest read on which side of that line a 4-to-20-person shop usually lands.
The Real Comparison Is a Stack, Not a Tool
A small commercial real estate firm does not own one proptech subscription. It owns a stack: a market-data service, a CRM, a listing tool, maybe a deal-management platform, an e-sign account, cloud storage, and a few point tools someone signed up for and forgot. The all-in number lives in your card statement, not on any one invoice.
That is the first correction to make before you touch a calculator. Generic “buy versus build” guides compare one SaaS product to one custom build, because they are written by software vendors for companies with engineering teams. Your decision is different: you are weighing a fixed automation project against the portion of the stack it would replace, plus everything that stack costs you in fees you keep paying and capability you never use.
The principle that decides it appears in our buy-versus-build playbook for CRE firms: standard workflows belong to vendors, firm-specific workflows are where a build earns its price. The three-year math puts a dollar figure on that.
What Three Years of Proptech Subscriptions Actually Costs
The subscription number you should compare is not the monthly sticker. It is the three-year total of every fee, every seat you add, every renewal increase, and the value lost to tools nobody logs into.
Four line items make up the real figure:
- Base subscriptions. A realistic all-in stack for a small brokerage, once market data is included, commonly runs from a few hundred to several thousand dollars a month. Market data is usually the heaviest line by far; the rest is CRM, marketing, and point tools.
- Per-seat creep. Most of the stack prices per user. Hire two analysts and the bill rises with headcount, not usage. Grow into a second market and a submarket-licensed data product can multiply, because you pay for coverage, not transactions.
- Renewal inflation. Proptech budgets at mid-market operators are rising an estimated 15 to 25 percent for 2026, much of it from AI features arriving as add-on SKUs rather than seat growth. Your stack does not hold its Year-1 price; model it going up.
- Shelfware. Zylo’s 2025 SaaS Management Index found roughly 53 percent of SaaS licenses go underused in a given month, and Gartner has estimated the average enterprise wastes about 30 percent of its SaaS budget on unused or duplicate tools. A slice of your subscription cost is buying capability that never touches a deal.
Put those together and the subscription case is not a flat line. It starts low and climbs, and the sticker undersells the three-year total once integration and the hours spent wrangling tools that do not talk are counted.
What a Custom Automation Project Actually Costs Over Three Years
A custom automation project is the opposite cost shape. It is expensive at the front and close to flat after.
Three components:
- The build. A single, well-scoped automation, one workflow such as turning a broker-blast inbox into a ranked pipeline or generating a first-draft investor report, generally lands between $25,000 and $150,000 depending on scope, integrations, and how messy the source data is. Most first projects for a small shop sit in the lower half of that band.
- The maintenance tail. Software is not a one-time purchase. Model prices shift, a vendor changes an export format, your process evolves. Budget an annual maintenance figure, commonly 15 to 20 percent of the build, to keep it current. For a firm with no IT department, this is the number that decides the whole thing, because you are buying someone’s ongoing attention, not just their code.
- Adoption. A build only pays back if the team uses it: a short training period and a few weeks of correcting early mistakes before it runs clean.
Notice what the build does not do: it rarely replaces the whole stack. You still pay for market data, e-sign, and storage. The build replaces the one workflow no vendor fits well, the one specific to how your firm makes money. So the honest comparison is the project against the slice of subscriptions it retires, not your entire software budget, a maintenance reality explored in our breakdown of what custom investor-reporting automation costs.
The 3-Year Math Side by Side
The figures below are illustrative patterns, not quotes; run your own numbers with the worksheet further down.
| Cost line | Proptech subscription (the slice you’d replace) | Custom automation project |
|---|---|---|
| Year 1 | Lower: base fees + setup | Higher: build + adoption |
| Year 2 | Rising: renewals up, seats added | Low: maintenance tail only |
| Year 3 | Higher still: compounding creep | Low: maintenance tail only |
| Cost curve | Starts low, climbs every year | Starts high, flattens |
| Scales with | Headcount and markets | Infrastructure only |
| You own | A license you rent | The workflow itself |
| Biggest risk | Price increases you can’t control | A maintenance tail you must fund |
The pattern is the whole point. Subscriptions win Year 1 on price and lose ground every year after, because per-seat pricing scales with your growth; a build costs more up front and then barely moves, because its marginal cost of another user is near zero. Independent total-cost-of-ownership analyses of SaaS versus custom software land on the same crossover repeatedly: the subscription is cheaper until it isn’t, usually somewhere in Year 2 to Year 3.
Where the Two Lines Cross
The crossover is the month cumulative subscription spend on the replaceable slice passes the cumulative cost of the build plus its maintenance. Three variables move it.
Headcount trajectory. The faster you add seats, the sooner subscriptions cross the build. A firm holding steady at six people may never cross; a firm going from six to fifteen crosses fast, because the per-seat stack grows with the team while the build does not.
How much the build retires. If the automation kills a $1,500-a-month workflow, it pays back quickly. If it only shaves an hour off a tool you keep anyway, the crossover recedes past the horizon.
The maintenance tail. Underfund it and the “cheaper after Year 1” story collapses, because a neglected automation decays into something the team stops trusting inside a quarter.
For most 4-to-20-person firms, the honest read is this: if you are not growing headcount and the workflow is standard, the lines may never cross, so subscriptions are the rational choice. If you are scaling and the workflow is specific to your firm, they cross inside three years and the build is the cheaper and better answer. The math rewards specificity and growth, and punishes building something a vendor already does well.
When Proptech Subscriptions Win
Buy, and keep buying, when the conditions favor renting capability rather than owning it.
- The workflow is standard across the industry. Rent rolls, e-signature, listing syndication, market comps. Thousands of firms do these the same way, so a vendor has already built the best version and spread the cost across all of them. You will not out-build that economically.
- You have not proven the volume. If a workflow runs a few times a month, a subscription is a cheap way to find out whether it matters before you commit capital.
- The vendor carries the maintenance. With no IT department, a subscription that updates itself is worth a premium over a build you must fund to keep alive. The choice to run your own data layer instead turns on this same maintenance question, covered in when CoStar is enough and when you need your own data.
- The all-in price is genuinely low. Some tools cost less per year than a single day of a developer’s time. Pay the fee and move on.
When a Custom Automation Project Wins
Build when the workflow is both expensive and yours.
- The workflow crosses systems that do not talk. When a task means copying data by hand between your CRM, a spreadsheet, and email because no vendor bridges them, that seam is a build candidate. Integration is where a custom project earns its price and where off-the-shelf tools quietly cost you hours.
- The workflow is how you make money. A generic tool makes you as capable as every firm that bought it. Automating the process specific to your edge, your underwriting method, your reporting format, is the one place software creates advantage rather than parity.
- Per-seat pricing has become a tax. Once a subscription’s cost scales faster than the value you get from it, usually because you have grown, a build with no per-seat fee overtakes it. The same math drives the choice in our look at deal-pipeline automation for boutique investment shops.
- Confidential deal data cannot leave your control. When the workflow handles data you would rather not feed into a shared vendor model, owning the pipeline is a governance decision as much as a cost one.
The Third Column Most Firms Should Run First
The decision is rarely a clean two-way fork, and the cheapest first move is usually neither a new subscription nor a build. It is fluency with the tools you already pay for. Much of what firms rush to automate, first-pass lease summaries, market write-ups, LOI drafts, cleaning up an inbox, is a well-built prompt over ChatGPT, Claude, or Gemini plus the spreadsheet you already run. A principal who gets fluent at that captures most of the value with zero new software, and learns which workflows are worth automating and which were never the bottleneck. That sequencing, fluency before spend, is the backbone of our training playbook for small CRE firms.
Fluency sharpens both other columns. It lowers the subscription case, because a fluent team buys less shelfware, and it sharpens the build case, because you can only write a good automation spec once you have done the workflow by hand. A short, hands-on workshop that teaches the team to prompt against real CRE tasks is the cheapest way to make a smarter buy-or-build decision later.
Run the Numbers Yourself Before You Sign
You do not need a spreadsheet model. You need six honest figures, in this order.
- The replaceable slice. Add up only the subscriptions a build would actually retire, not your whole stack. This is what the build competes with.
- Three-year subscription total. Take that slice, add per-seat growth for your realistic hiring plan, and inflate renewals by roughly 15 to 20 percent a year.
- Shelfware discount. Subtract the tools nobody truly uses. Dead licenses are cost with no value, and they strengthen the case to consolidate or build.
- Build estimate. A scoped range for the one workflow, plus annual maintenance at 15 to 20 percent. Three years is build plus two years of maintenance.
- The crossover. Compare the two totals. If the build is lower and also frees hours, build; if it is higher, subscribe unless data control forces the build.
- The fluency test. Before either, ask whether a fluent team could get most of the value with the tools you already own. If yes, do that first.
Firms that skip step one and compare a build against their entire software budget always over-buy the build. Firms that skip step six always over-buy software. The discipline of counting real volume before committing runs through our math on automated market reports versus hiring an analyst, and it is the logic that lets a lean shop out-operate a bigger one, as argued in the small-firm CRE manifesto.
Frequently Asked Questions
Is it cheaper to subscribe to proptech or build custom automation?
It depends on your growth and the workflow. In Year 1, subscriptions almost always win because a build costs its full price up front. Over three years the picture flips for firms adding headcount or automating a firm-specific workflow, because per-seat fees compound while a build stays roughly flat. Compare the three-year total of only the subscriptions a build would replace against the build plus its maintenance. If you are not growing and the workflow is standard, subscriptions usually stay cheaper the whole time.
How much does a custom automation project cost for a small CRE firm?
A single, well-scoped workflow automation generally falls between $25,000 and $150,000, most first projects in the lower half depending on scope, integrations, and how messy the source data is. Budget annual maintenance of roughly 15 to 20 percent of the build to keep it current. For a firm with no IT department, that maintenance line matters more than the build price, because you are funding someone’s ongoing attention, not a one-time delivery.
What does a proptech subscription really cost over three years?
Far more than the monthly sticker. The three-year total includes base fees, per-seat increases as you hire, renewal inflation running an estimated 15 to 25 percent for 2026 at many operators, and value lost to shelfware. Zylo’s 2025 index found roughly half of SaaS licenses go underused in a given month, and Gartner has estimated the average enterprise wastes about 30 percent of its SaaS budget. Model the subscription side as a rising line, and count only the portion a build would replace.
When does the cost of subscriptions overtake a build?
Usually somewhere in Year 2 to Year 3, if it happens at all. It comes sooner when you add seats quickly, when the build retires an expensive workflow rather than a cheap one, and when maintenance is funded. It never comes for a firm that stays the same size and automates a standard workflow. Headcount trajectory is the single biggest driver of where the lines cross, so map your hiring plan first.
Can we replace our whole proptech stack with one custom build?
Almost never, and be wary of any vendor who says you can. A custom automation replaces the one workflow specific to your firm and poorly served by off-the-shelf tools. You will still pay for standard capabilities like market data, e-signature, and listing syndication, because building those costs more than renting the best version. Think of a build as retiring a slice of the stack, and compare its cost only against that slice.
What if we grow from six people to fifteen?
Growth is the strongest argument for building. Per-seat pricing scales with headcount, so a stack that was affordable at six can become a real line item at fifteen, while a custom automation costs nearly the same to run for either team size. If you have a credible plan to scale, model the subscription side at your future headcount, not today’s, because that is the cost you will actually carry.
Should a small firm with no IT department build at all?
Yes, but only with a maintenance plan in place first. No IT department does not rule out a build; it changes what you are buying: a delivered automation plus a commitment from whoever built it to keep it running. Never commission a project without a clear answer to who maintains it and what that costs each year, because an unmaintained automation decays into something the team abandons within a quarter.
Is it worth building if the software already exists off the shelf?
No. If a vendor already does the workflow well, building your own version makes you as capable as everyone who bought that vendor, at higher cost and with a maintenance burden you now own. Building pays only when the workflow is specific to how your firm makes money or crosses systems no vendor bridges. Reserve the build budget for the workflow that is genuinely yours, and rent everything standard.
Where to Start
The first move is not to pick a vendor or approve a build. It is to separate the slice of your stack a build could actually replace from the industry-standard tools you will keep renting, then decide whether a fluent team could get most of the value with the software you already pay for. Answer those two questions and the three-year math mostly resolves itself. A free AI-readiness assessment gives you that read: a short working session that looks at your stack, your workflows, and where your hours actually go, then returns an honest recommendation on whether to buy, build, or get fluent first. Book a free AI-readiness assessment before you renew a stack or approve a project you may not need.
Arthur Wandzel
