Productising an Agency Service: How to Turn Your Best Engagement Into a SKU

Every agency has that one engagement they've delivered seven times. Same shape, same deliverables, same objections from the client, same three weeks of work. And every time, someone on the sales side writes a fresh proposal from scratch and quotes a slightly different number.

That engagement is a product waiting to happen. Not a SaaS — a productised service. And in our experience it's a far better first step out of pure time-and-materials than trying to build a side-project SaaS while running an agency.

Why productise before you try to build a product

The agency-to-product jump is romanticised. Founders read about Basecamp and ConvertKit and assume the move is: keep the agency running, moonlight a SaaS, flip the ratio over 18 months. In reality, most of those attempts die because the agency's cash flow needs attention every single week and the SaaS never gets a full-time owner.

A productised service sits in between. It's still services revenue, so it funds itself. But it behaves like a product in the ways that matter:

• Fixed scope, fixed price, fixed timeline
• One landing page, one checkout conversation, one onboarding flow
• A delivery playbook that a mid-level engineer can run without the founder in the room
• Margin that improves with each delivery instead of degrading

If you can't run a productised service profitably, you almost certainly can't run a SaaS. The discipline is the same: narrow the ICP, kill scope creep, and measure delivery cost obsessively.

The candidate engagement

Look at your last 18 months of invoices. You're hunting for an engagement that meets four criteria:

1. You've delivered it at least four times
2. The scope barely changed between deliveries
3. Clients came to you asking for roughly the same outcome
4. Your gross margin was healthy — north of 40% — even without optimising

Common candidates we see: Shopify Plus migrations for mid-market retailers, a 6-week AI feasibility audit, a Next.js rebuild of a legacy WordPress marketing site, a HIPAA-adjacent readiness review for a Series A health startup, a Stripe billing integration for a B2B SaaS moving off manual invoicing.

If nothing on your list fits, don't force it. Productising a one-off is how you invent a product no one wants.

Scoping the SKU

This is where most attempts fall apart. Founders write a page of marketing copy, slap a price on it, and call it productised. Then the first buyer asks a reasonable question — "does this include the mobile app?" — and the whole thing collapses back into custom work.

A productised service needs three documents before you sell it once.

The deliverables list

Write down every artefact the client receives. Be boring and specific. Not "technical audit" but "a 20–30 page PDF audit covering the following eight areas, plus a 90-minute walkthrough call with the CTO."

The exclusions list

This one matters more than the deliverables list. What is explicitly not in scope? We include this on the sales page, not buried in the SOW. Buyers respect it and it filters out the wrong clients before they hit your calendar.

The delivery runbook

Internal document. Day-by-day or week-by-week, what happens, who does it, what tools they use, what the handoff looks like. If you can't write this, you don't have a product yet — you have a talented senior consultant.

A rough shape for a 6-week audit engagement:

sku: ai-feasibility-audit
duration_weeks: 6
team:
  - role: lead_engineer
    allocation: 0.5
  - role: solutions_architect
    allocation: 0.25
  - role: pm
    allocation: 0.15
phases:
  week_1: kickoff, data access, stakeholder interviews
  week_2_3: technical discovery, model shortlisting
  week_4: prototype on client data
  week_5: cost modelling, risk register
  week_6: written report, exec readout
deliverables:
  - discovery_notes.md
  - prototype_repo
  - cost_model.xlsx
  - final_report.pdf
  - exec_readout_deck
excluded:
  - production deployment
  - ongoing model monitoring
  - data pipeline construction

The runbook is what lets you delegate delivery. Without it, the SKU is a founder-shaped bottleneck.

Pricing without guessing

The temptation is to price a productised service by taking your blended day rate, multiplying by estimated days, and rounding up. That gives you a services price, not a product price.

Instead, price against three anchors:

Cost floor. Fully-loaded delivery cost including PM overhead, tooling, and a realistic buffer for the one client per quarter who eats extra hours. If your gross margin drops below 50% at this price, the SKU isn't productised enough yet.

Value ceiling. What's the buyer's alternative? If they'd otherwise hire a full-time senior engineer for three months at $60k loaded cost, you're not selling a $12k audit — you're selling $60k of avoided risk and time.

Market anchor. What do buyers already pay for adjacent things? Nobody buys in a vacuum. If comparable audits from Big 4 consultancies land at $80–120k and boutique shops at $25–50k, you know the corridor.

We usually land in the top third of the boutique corridor, with a clear reason why. Discounting on a productised service is poison — the moment you negotiate on price, you signal that scope is negotiable too, and you're back in T&M land.

The good-better-best trap

Everyone reads about tiered pricing and tries to ship three variants on day one. Don't. Ship one SKU, sell it 8–10 times, then decide whether to add a lighter or heavier tier. Most of the time you'll discover the market wants a completely different second SKU, not a variant of the first.

Selling it without confusing the buyer

A productised service sits awkwardly in the sales conversation. Buyers are conditioned to expect either a SaaS signup or a bespoke proposal. When you offer a fixed-scope, fixed-price engagement, some buyers relax immediately and others get suspicious.

A few things that have worked for us:

• A dedicated landing page, separate from your main services pages. Treat it like a product page: hero, deliverables, timeline, price, FAQ, buy-now CTA (even if the CTA is "book a 20-minute fit call").
• Publish the price. Half the point of productising is filtering. Hiding the price re-invents the RFP dance you're trying to escape.
• A fit call, not a sales call. 20 minutes, one goal: decide whether this SKU is the right fit. If it's not, refer them to your custom services page or a partner. This builds enormous trust.
• A written SOW that mirrors the landing page exactly. Same deliverables, same exclusions, same timeline. If the SOW starts drifting from the marketing page, the SKU is decaying.

We link the fit-call booking straight from our services pages so the buyer's path is obvious.

The traps that kill margin

Three failure modes we've watched destroy otherwise-good productised services:

Scope leakage. Client asks for "just one small extra thing" in week 3. You say yes because the relationship matters. By week 6 you've done 15% more work for the same fee. Fix: any change request goes into a separate mini-SOW at your standard T&M rate, no exceptions. Not even for good clients. Especially not for good clients — they'll respect the boundary.

Senior-person creep. The runbook says a mid-level engineer runs delivery. In practice the founder joins every client call because it feels safer. Margin evaporates. Fix: the founder joins the kickoff and the final readout. Nothing in between, unless something is actually on fire.

Custom onboarding. Every client gets a bespoke Notion workspace, custom Slack channel structure, tailored kickoff deck. Fix: one template, one folder structure, one deck with the client's name swapped in. Boring is the point.

When to graduate the SKU

After 15–20 deliveries you'll know whether the SKU has legs. Signs it's ready to evolve:

• Delivery cost is trending down, not up
• You're getting inbound specifically for the SKU, by name
• Clients are asking "what's next?" after delivery — a natural retainer or phase two
• Your team can run it without you

That's when you start thinking about the second SKU, or about the actual product — the SaaS that automates the most expensive part of the delivery. The runbook you wrote in month one becomes the product spec in year two.

Where we'd start

If you're running an agency and this resonates, do one thing this week: pull your last 18 months of invoices into a spreadsheet, tag each engagement by shape, and find the row that repeats most often with the healthiest margin. Don't design a SKU yet. Just look at what your business is already telling you it wants to sell. The productisation work is mostly listening — writing down what you're already doing well, and refusing to do the parts you're not.

Everything after that is discipline. Which, honestly, is the whole game.