Custom App vs. Off-the-Shelf Software: How to Decide

Every week, someone asks me some flavor of: "should we build something custom, or is there software that already does this?"

The honest answer is almost always "there's software that already does this, and you should buy it first." Custom software is a last resort, not a default. But there's a specific set of conditions where custom pays off dramatically — and I see a lot of businesses either commit to custom too early (and overspend by 10x) or stay on off-the-shelf too long (and overspend on licensing forever).

Here's the framework I walk through on every Discovery Call.

The default: buy, don't build

Assume off-the-shelf wins, until you prove otherwise. Here's why the math favors buying:

• Someone else already spent $5M+ building it
• You pay a fraction of that as a subscription
• They keep improving it whether you ask or not
• If you hate it, you can switch
• Your team probably already knows how to use something in the category

Building custom flips all of those. You pay the full dev cost. You own maintenance forever. You're the only person improving it. Switching means throwing it away.

So before we even get to "should you build custom," the real question is: can you make off-the-shelf work?

The four questions

I ask these four questions in order. If you get a clean "yes" on all four, build custom. If any one is a "no" or "unclear," start with off-the-shelf.

Question 1: Is your workflow stable?

Custom software is built around your specific workflow. If your workflow is going to change significantly in the next 12-18 months — because you're still figuring out your business, because you're about to acquire a competitor, because your market is shifting — don't build custom. The code will be obsolete before it earns its keep.

Off-the-shelf tools are designed to be generic on purpose. If your workflow is in flux, generic is an asset. You can bend generic tools to fit whatever shape your business takes this quarter.

Question 2: Is the work you do high-volume and repetitive?

Custom pays off when the same process happens over and over. One-off work? Use a tool. Ten times a year? Use a tool. Ten times a day? Now we're talking.

Rule of thumb: custom earns its keep when the process runs at least 3x per week, for at least 12 months forward. Below that threshold, you can't save enough human hours to pay for the build.

A property management company processing 40 move-outs a month = worth automating. A consulting firm managing 3 complex engagements a quarter = not worth automating.

Question 3: Does any off-the-shelf tool genuinely cover what you need?

The trap is that off-the-shelf always almost covers it. HubSpot almost fits. Salesforce almost fits. AppFolio almost fits. Airtable almost fits. Monday almost fits.

The question isn't "does something exist that's in the same category." The question is: does any existing tool cover 80%+ of what you need without you having to hack around it?

Signs you're hacking around a tool:

• You maintain a "workaround doc" to explain how your team uses it
• You have a column called "notes" that holds structured data the tool doesn't support
• You export data regularly to work on it somewhere else, then re-import
• You have a dedicated person whose job is partially "keep the tool happy"
• You pay for seats for people who barely use the tool just to preserve access

If any two of those are true, you've outgrown the tool. You're now paying full price for 50% utility.

“

We were paying $18,000/year for Salesforce and the real truth was that three of us used it for one thing. The other seats were historical.

Question 4: Is the cost of the off-the-shelf option (over 3 years) more than the build cost?

Run the math. Seriously, run it.

Off-the-shelf 3-year cost:

• Monthly license × 12 × 3
• Plus implementation / onboarding costs
• Plus integration maintenance (often 10-20% of license cost)
• Plus seats for people who only need partial access

Custom 3-year cost:

• One-time build cost
• Hosting (usually $40-$200/mo)
• Maintenance: assume 15% of build cost per year
• Any new features over that span

For the property management company I worked with, the HubSpot-plus-middleware stack was costing $24K/year. Custom build was $18K one-time, $2K/year in maintenance. Over 3 years, HubSpot was $72K. Custom was $24K. Over 5 years, $120K vs $28K.

That's not a marginal difference. That's a real dollar figure.

The middle path you probably didn't consider

There's a third option nobody talks about: buy the core, build the edges.

This is where most SMBs should actually land. You keep an off-the-shelf tool for the 80% that's generic (CRM data, invoicing, calendar, file storage), and you custom-build only the specific workflow pieces that don't fit.

For the PM company, this looked like: keep AppFolio (it's legitimately good at tenant/unit management, and building that from scratch would be insane). Replace HubSpot with custom pipeline tools that pull from AppFolio.

The wins:

• You don't have to rebuild all of AppFolio's core functionality
• Your team keeps using the tool they already know for the boring parts
• Your custom code focuses on the parts that actually matter
• Build cost drops by 60-80% compared to full custom

This is the model I use for almost every client now. Pick the right off-the-shelf core, then custom-build the thin layer of workflow that matters.

The sign you've gone custom too soon

If you find yourself building features in your custom app that you could literally get for free in Notion, you went custom too early.

Examples of features that are almost never worth custom-building:

• Markdown editors (use an existing one)
• File uploads with thumbnails (cloud providers do this)
• Calendar views (too many libraries to count)
• Login / auth (use Auth0, Clerk, or Supabase — never roll your own in 2026)
• Email sending (use Resend, Postmark, or SendGrid)
• Payment processing (Stripe)

Custom code should focus on the specific business logic that makes your company different. Everything else should be glued-together off-the-shelf pieces. If your dev is "building auth" from scratch, that's $10K of work with no business value.

The sign you've stayed off-the-shelf too long

You know you've stayed off-the-shelf too long when:

• Your team says things like "we just do that manually because the system can't handle it"
• Onboarding a new hire takes 2+ weeks because your tool stack is a Rube Goldberg machine
• You're paying for three tools that partially overlap because none of them cover the full workflow
• Your IT / ops person spends most of their time being glue between tools

At that point, the cost of not building custom is showing up as salary and salary-adjacent costs instead of software subscriptions. It's easier to hide that cost, but it's still real.

The actual decision

Here's the simple heuristic I use to recommend a direction after a Discovery Call:

1. Is there a well-known tool in the category? If no, you probably have a truly unique workflow and custom makes sense. If yes, try that first.
2. Are you at least 18 months into doing this work the same way? If no, stay flexible with off-the-shelf. If yes, you can commit to custom.
3. Does your annual spend on partial-fit tools exceed $10K? If yes, custom is probably cheaper over 3 years. If no, stay bought.
4. Can you tolerate a 6-12 week build window with no new features? If no, you're not ready for custom.

Four yes's = build. Any no = buy, audit in a year.

---

If you want me to run this framework on your specific situation — with real numbers, no pitch — that's what the Discovery Call is for. 15 minutes. I'll tell you what I'd actually do in your shoes.