Can a Small Business Actually Afford an FDE?

The owner of a 12-person manufacturing shop sat across from us at a diner and asked the question straight. "Can a small business actually afford a forward deployed engineer? Or is this only for companies with a hundred people and a CIO?" He had a spreadsheet open on his laptop showing six hours of weekly invoicing chaos, three different inventory lists that never matched, and a quoting process that took two days because nobody trusted the numbers. Hiring a forward deployed engineer for small business ops can work if the pain is specific and the budget is honest.

Our answer was yes, with conditions. A small business can afford an FDE if the scope is tight, the funding comes from savings on the current broken process, and everyone agrees that the first build is a test, not a transformation. The role was born at Palantir scale, but the model shrinks well. A 30-person company with one expensive manual workflow can get more value from a two-week embedded engagement than a 300-person company gets from a six-month consulting contract.

The key is specificity. If you do not have a specific painful process with a measurable cost, you are not ready for an FDE. You are buying hope, not engineering. And hope is expensive.

What Small Business Engagements Look Like

Small business FDE engagements are shorter, narrower, and more focused than enterprise ones. Instead of redesigning operations, you fix one workflow. Instead of a team of stakeholders, you have one decision-maker. Instead of a six-month roadmap, you have a two-week sprint with a clear finish line.

A typical small engagement looks like this. Last month we shadowed a bookkeeper who spent six hours every Friday manually reconciling Stripe payouts against invoices. Week one: shadow the person doing the painful work, document the exact steps, and agree on a replacement workflow. Week two: build a minimal tool that handles 80% of the volume, train the user, and watch it run for three days. Week three: fix the bugs that only appear in real use, add the one missing report everyone forgot to ask for, and hand it off.

The total cost? A focused two-week pilot typically runs $8,000 to $15,000. A full build with integration might be $25,000 to $50,000. That sounds like a lot for a small business until you compare it to the cost of the status quo.

The Self-Funding Pilot

A better way to pay for an FDE is with the money you are already losing to the broken process. We call it the self-funding pilot. You calculate the annual cost of the manual work, the error rate, and the revenue you cannot capture because the process is too slow. If that number is higher than the build cost, the math works.

Say you are a 10-person firm and one employee spends six hours a week on manual invoicing. At $30 an hour loaded cost, that is $9,360 a year in labor. Add the cost of errors, late payments, and the quotes you lose because they take two days to produce. The total is probably $20,000 to $30,000 annually. An $18,000 build that cuts that work by 80% pays for itself in under a year.

The mistake small businesses make is comparing the FDE cost to zero. The right comparison is the FDE cost to the current process cost. Manual work is not free. It is just invisible on the balance sheet.

Red Flags: When an FDE Is Too Much

Not every small business needs an embedded engineer. Here are the red flags that tell us to walk away.

If the owner says "we just need something better" without naming the workflow, the scope is too vague. If the team is five people and nobody spends more than an hour a week on the target process, the savings will not justify the cost. If the business is in crisis, cash flow is unpredictable, and the owner is hoping software will save the company, the timing is wrong. Software does not fix a broken business model. It amplifies whatever is already there.

Another red flag: no one internally can own the tool after the FDE leaves. Small businesses do not have IT departments. If the owner is too busy to learn the basics of the new system and there is no one else to maintain it, the tool will rot. We always ask who will handle updates, answer user questions, and decide when something needs changing. If the answer is crickets, we decline.

Alternatives Before an FDE

Sometimes the right answer is not an FDE. It is Retool, n8n, or a good freelancer. These options get you 70% of the way there for 30% of the cost.

Retool is excellent for internal dashboards that connect to existing databases. If your problem is "we cannot see our data in one place," Retool might solve it in a week without custom code. n8n handles workflow automation between SaaS tools. If your pain is "when a sale happens in Stripe, someone has to manually create an invoice in QuickBooks," n8n can bridge that gap for a few hundred dollars a year.

A skilled freelance developer can also build a focused tool for less than an FDE engagement, though you lose the embedded discovery process. The freelancer builds what you ask for. The FDE figures out what you actually need. That difference matters when you are not sure where the pain originates.

We usually tell small businesses to try the off-the-shelf route first. If it works, great. If it gets you to 70% and the remaining 30% is costing you real money, that is when an FDE makes sense. You are not starting from zero. You are sharpening a tool that already proved its value.

Making the Decision

We use a simple go/no-go framework based on three numbers. Annual process cost, team size, and pain specificity.

Calculate the annual cost of the process you want to replace. Include labor, error correction, and opportunity cost. If it is under $15,000 a year, an FDE is probably overkill. If it is over $30,000, the math is compelling. Next, team size. A team of 5 to 10 with one process costing 10-plus hours weekly is usually viable. Below five people, the coordination overhead of embedding an engineer can exceed the savings. Above twenty, you might need a longer engagement with more stakeholders.

Finally, pain specificity. Can you describe the problem in one sentence that includes a dollar amount? "Our bookkeeper spends six hours a week reconciling two systems that should talk to each other" is specific. "We need to be more digital" is not. The more specific the pain, the tighter the scope, the lower the risk, and the faster the ROI.

Small businesses do not have room for expensive experiments. They need tools that pay for themselves quickly and keep working without a dedicated IT team. A forward deployed engineer can deliver that, but only when the conditions are right. Tight scope, specific pain, and savings that fund the work. Get those three right, and a 12-person shop can out-automate a competitor ten times its size.

If you are wondering what it actually feels like to work with an embedded engineer, our post on what a forward deployed engineer does breaks down the day-to-day. For the client's perspective, what it's like to work with an FDE explains the relationship from the other side of the table. And if you want to know why clients keep us around, why clients keep FDEs around covers the long-term view. The model works at scale, but it works just as well at a diner table with a laptop and a specific problem.