What SMBs actually need from custom software

title: "What SMBs actually need from custom software" description: "After 40+ small-business engagements, the pattern is obvious: it's never the cool feature, it's the boring workflow. A field guide to what's worth building." date: "2026-05-22" tags: ["smb", "product", "strategy"] author: "PremiumWebApps Studio"

When a small-business owner says "we need an app," they almost never mean an app. They mean: there's a workflow that's eating my week, and I don't have a way to fix it.

Here's what those workflows actually look like, ranked by how often we see them.

1. The spreadsheet that grew up

The single most common engagement we take is: "We've been running [thing] in Excel for years and it's broken." It's never broken because Excel is bad. It's broken because:

• Multiple people need to edit simultaneously and they're sending versions over email.
• There's no audit trail when something goes wrong.
• The formulas are too fragile — one wrong paste breaks everything.
• It can't talk to anything else (Stripe, QuickBooks, the warehouse system).

The fix is almost never "build us Excel, but better." It's: identify the three actions people actually do, build clean UI for those, and let the rest stay in Excel forever.

2. The email-and-PDF workflow

Quoted, then invoiced, then approved, then signed, then filed, then reconciled — all over email with PDFs. Every SMB has one of these. The cost isn't the time it takes; it's the leaks. Quotes get lost. Approvals don't happen. Things slip past 60 days.

We replace these with a tiny portal: one URL per customer, one page per stage, status visible to everyone. Not a CRM. Not a full ERP. Just the specific flow that's leaking.

3. The "we want an app like X"

The owner saw a slick app from a $100M company and wants the same thing for their 40-person business. This is the engagement we say no to most often, because the cool feature isn't what's making that company money — it's the unsexy operations underneath. We try to reroute the conversation toward the workflow that's actually broken.

When the cool-app pitch does make sense, it's usually:

• Customer-facing booking or ordering (where the friction loses revenue daily).
• Field-team data capture (where the existing tool is a PDF on a clipboard).
• A marketplace where they're stuck in the middle (matchmaker, brokers, agencies).

4. The AI add-on to an existing thing

Increasingly common: "We already have [legacy system / website / internal tool]. Can you bolt AI on?" Almost always yes, almost always cheaper than they expect, almost always the easiest project we'll take that month.

The usual shapes:

• Document Q&A over their existing files (legal, real estate, technical libraries).
• Customer-support deflection trained on their actual policies and history.
• Internal-search that finally finds things in their Notion / SharePoint / Drive.
• Smart routing of incoming leads / tickets / orders based on rules + LLM judgment.

You don't rebuild the whole stack to add AI. You build a thin layer that reads the right data, calls a model, and writes back somewhere useful.

5. The internal tool no one asked for

A specific person at the company is doing 4 hours/week of clicky-clicky between three SaaS tools. The owner doesn't notice. The bookkeeper, the warehouse lead, the head of HR — they know exactly where the bleeding is. Ask them.

These engagements are tiny (often $4k–$6k) and the ROI is brutal. We've shipped 8-hour-payback projects for clients twice in the last quarter.

What this means for your project

Before you scope anything, answer these three questions. The cost of not answering them is at least 2× the build cost in rework:

1. Who specifically uses this, every day? Name them.
2. What three actions do they perform with it? No more than three.
3. What's the cost — in time or dollars — of the current workaround? Per week.

If you can answer all three in plain English, you're ready to scope. If you can't, we'll help on the discovery call.

Request a quote →