For a 10–200 person company the fastest, lowest‑risk route is a fixed‑scope consultant engagement that delivers one production automation with measurable acceptance criteria, a rollback plan, and a short warranty. Buy only when an off‑the‑shelf app matches the process exactly; build or hire only after repeatable demand is proven.
- I usually recommend a fixed‑scope consultant engagement to deliver one reliable automation into production fast for teams of 10–200 people.
- Buy an off‑the‑shelf app only when the process is simple, stable, and maps exactly to the app's workflow.
- Hire a full‑time automation person or build a platform only after you can justify repeatable demand and ongoing headcount.
- Plan and budget ongoing operations; monitoring and fixes typically cost a portion of initial delivery every year.
- Define acceptance criteria and a rollback plan before any pilot; pilots without them become graveyard projects.
Signs this is happening in your business
- Three people on your team spend six hours each week copying CRM fields into invoices and chasing missing data.
- Month‑end takes several extra working days because someone assembles numbers into a spreadsheet manually.
- Support forwards tickets between systems and no one tracks which ones are waiting on finance.
- One person keeps every workaround documented in their head and leaving would break the process.
- You bought a point app but the team still uses a spreadsheet to glue workflows together.
- An automation stopped working after a supplier changed a form and nobody noticed.
- You have many pilot projects but none are in daily use; they are quiet on Mondays and forgotten by Thursday.
- Finance still emails PDFs that someone must retype into the accounting system.
- Reports are produced by a single person who refuses to let others touch the spreadsheet.
If several of these are true at the same time, the problem is usually the process rather than the people running it.
What this is costing you a year
(Team hours per week x hourly cost of employees x 52) = annual payroll loss
- People doing this work4
- Hours each per week6
- Hourly cost of employees$45
- Weeks52
The annual payroll cost across the team spent on this manual work.
Find the number before you spend anything
Run the formula above on your three most repetitive processes. If any of them clears a year of a salary, that is the one to bring to a call.
When to buy, build or hire automation for a 10–200 person company
For companies of 10–200 people, buy when a process is simple, stable and fits an app exactly; hire a fixed‑scope consultant when you can specify a single high‑cost process in a week and need production results fast; only hire full‑time or build a platform when demand is repeatable and the numbers support headcount.
- Buy an app when the process is simple and unchanged.
- Hire a consultant for one reliable production process delivered fast with limited risk.
- Build or hire full‑time only when you can prove repeatable demand and ongoing work.
How much each route costs in year one and over three years
I give worked total cost of ownership examples in USD for a single medium‑complex process and list the hidden costs you must include. Use these models to rerun the math with your inputs and to spot what pushes a cheap pilot into an expensive long‑term problem.
- Off‑the‑shelf app: license fees, configuration time, training, data migration, per‑user fees, and mismatch costs when the app doesn't match the process.
- Fixed‑scope consultant: delivery fee, knowledge transfer, warranty window, monitoring setup, and an optional maintenance retainer.
- Platform/custom build: design and development, hosting and monitoring, ongoing bug fixes, and staff time for ops and feature requests.
- Full‑time hire: salary and benefits, onboarding time, tooling, reduced velocity while they learn systems, and employer cost when they leave.
Example TCOs for one medium‑complex process
Below are three sample TCOs in USD for a medium‑complex process that connects two systems, needs data validation and occasional manual approvals. Each sum includes delivery, first‑year maintenance and the hidden costs I commonly see so you can swap in your own inputs.
- Off‑the‑shelf app (example): setup $6,000 first year; ongoing fees $2,000/year; configuration and training 80 hours at $60/hr = $4,800. Year‑one TCO ≈ $12,800. Three‑year TCO ≈ $16,800 plus switching costs if it fails to match.
- Fixed‑scope consultant (example): delivery fee $18,000 (includes acceptance tests and handover), monitoring setup $2,500, optional annual retainer $4,000. Year‑one TCO ≈ $20,500. Three‑year TCO with one annual retainer ≈ $28,500.
- Platform / custom build (example): design & development $80,000, hosting and monitoring $6,000/year, ongoing improvements 0.4 FTE ≈ $35,000/year. Year‑one TCO ≈ $121,000. Three‑year TCO ≈ $226,000.
- Full‑time hire (example): salary + benefits $140,000/year, tooling and training $6,000. Year‑one TCO ≈ $146,000. Three‑year TCO ≈ $438,000 ignoring turnover.
Automating a broken process only makes brokenness happen faster and cheaper; fix the process first.
Speed to production and common failure modes
Time from pilot to production depends on decision speed, defined acceptance criteria, and who owns handover. I list realistic timelines and the failure patterns I see: unclear acceptance, absent monitoring, and no operational owner after delivery — the three things that stop projects shipping.
- Off‑the‑shelf app: pilot to production 2–6 weeks; failure = abandoned configuration and the team keeps the spreadsheet.
- Fixed‑scope consultant: pilot to production 4–12 weeks; failure = incomplete handover and no one to own fixes.
- Platform/DIY build: pilot to production 3–9 months; failure = scope creep and no ops budget so it degrades.
- Full‑time hire: 3–9 months to full productivity; failure = single‑person dependence and competing firefighting work.
Decision thresholds and exact signals that justify a full‑time hire
Hire a full‑time automation person when repeatable demand and the maths justify headcount. Look for queued automations, total hours saved, frequent process change, and the need for platform capabilities — these are cashflow‑sensitive thresholds I use in scoping sessions.
- At least five medium‑complex automations queued or created every 12 months, each saving at least 20 staff hours per month.
- Total monthly hours saved exceed 160 hours and the dollar value of those hours exceeds the fully loaded cost of the hire.
- Frequent changes to processes or integrations that need ongoing development rather than one‑off delivery.
- You require platform capability: reusable components, dashboards, and internal governance for many automations.
Handover, ownership and avoiding post‑delivery technical debt
Decide ownership before any work starts. Name an ops owner and an escalation path or the automation will fail within weeks. A proper handover contains measurable acceptance checks, usable documentation, and a short warranty with defined response times.
- Define acceptance criteria as business checks (for example, 'every invoice processed with under 1% data error for 30 days').
- Contract a rollback plan so you can disable the automation and return to manual state without data loss.
- Include 30–90 days of monitoring and a defined time budget for fixes; expect maintenance to cost a portion of initial delivery in year one.
Pilots without a rollback plan are audits you will fail later; every production automation needs a shadow period.
Practical checklist to decide for each process
Score the process quickly on integration surface, variability, stakeholders, compliance and frequency to pick buy, consultant, or build. High complexity and frequent rule changes push you toward consultant or build; low complexity and stable rules favour buying or hiring a consultant for a fixed scope.
- Complexity: how many systems are involved? (1 system = buy possible; 3+ systems = consultant or build).
- Variability: how often do business rules change? (rare = buy; frequent = hire/build).
- Stakeholders: how many teams approve the process? (1–2 = buy/consultant; 3+ = consultant/build).
- Compliance/security: regulated data means avoid off‑the‑shelf unless vendor certified.
- Frequency: daily or more increases automation value.
- ROI threshold: monthly payroll hours saved must pay back delivery in 12–18 months for consultant to be the clear winner.
How this fails in practice — the sequence you will recognise
Most automation projects die in three stages: a pilot without acceptance, a handover without ops, and then silence. Spotting this sequence early lets you stop it by insisting on acceptance tests, monitoring and a named operational owner before you accept delivery.
- Stage 1 — Excitement: a small team builds a pilot with no acceptance tests and it passes in the demo.
- Stage 2 — Delivery hangover: no documentation, no monitoring, edge cases undocumented; the pilot breaks when something changes.
- Stage 3 — Graveyard: no on‑call, team reverts to spreadsheets, and the build budget is treated as sunk.
Identify your biggest operational bottleneck
Most owners already know which process is the problem. What they do not have is the annual figure attached to it. That is a 15-minute conversation, not a project.
Two worked examples that make the choice obvious
Two examples—CRM re‑keying and month‑end reporting—show the arithmetic you need to choose buy, hire a consultant, or justify a headcount. I keep the numbers visible so you can swap in your own rates and run the payback calculation.
- CRM re‑keying inputs (USD): team = 3 people, hours each per week = 5, hourly cost = $40, weeks = 52. Annual payroll loss = 3 × 5 × $40 × 52 = $31,200. Off‑the‑shelf app cost: $9,000 first year. Consultant delivery: $14,000 with 6 months warranty. Build/headcount: $130,000 hire yearly.
- Month‑end reporting inputs (USD): team = 4 people, hours each per week = 8 on average, hourly cost = $50, weeks = 52. Annual payroll lost = 4 × 8 × $50 × 52 = $83,200. A consultant who automates the pack for $30,000 with a $5,000 annual retainer pays back in under a year; build only if you plan many such automations.
Hiring a single automation person before you have repeatable demand wastes salary and creates a single point of failure.
Who should own maintenance, SLAs and runbooks
Split ownership between a business owner who measures outcomes and a technical ops owner who handles alerts, runbooks and safe disablement. Without both, automations decay. Write SLAs in business terms: acknowledge time, time to fix or rollback, and escalation rules.
- Business owner: owns the metric (for example, 'invoices processed per day') and the acceptance criteria.
- Ops owner: owns monitoring alerts, runbooks, and the ability to disable the automation safely.
- SLA example: acknowledge in 4 business hours; fix or rollback in 2 business days for production failures.
How to contract a fixed‑scope delivery so it actually lands
Contract a fixed scope with explicit acceptance criteria, a rollback plan, a warranty, knowledge transfer sessions, and an optional short maintenance retainer. Require runbooks, a list of edge cases, and a shadow period where the manual process runs in parallel while the automation is verified.
- Acceptance criteria must be measurable business checks, not 'works most of the time'.
- A 30–90 day shadow window prevents surprises when the automation runs at scale.
- Include a short retainer rather than indefinite free support to reduce long‑term risk.
If you need leadership rather than a build person
If you lack technical leadership, hire an experienced fractional CTO or adviser to prioritise which processes to automate and to design a repeatable delivery pipeline. A tactical hire without that leadership will produce a couple of automations but no durable capability.
- I provide advisory work under Tech consulting and can help set acceptance criteria and governance.
- A fractional leader prevents single‑person dependencies and keeps the automation roadmap aligned with commercial goals.
Why most rival articles mislead you
Many guides list pros and cons and avoid the money and operations. I focus on cash maths, acceptance criteria, monitoring and handover because pilots without those elements are demos, not production outcomes.
- Don't be seduced by cheap pilots that have no ops plan.
- Don't value a demo's speed over sustained daily reliability.
Next step: scope one high‑cost process as a fixed‑price pilot
Scope a single high‑cost process into a fixed‑price pilot with clear acceptance criteria and a short warranty. That low‑risk step proves the numbers faster than buying an app or hiring a person first and reduces the chance of creating another abandoned proof‑of‑concept.
- A scoping call gives you an estimate, acceptance criteria and a payback calculation you can use to fund delivery; book via How to work together.
- See examples and past work at Portfolio and case studies to compare problems and outcomes.
Resources and services I use to deliver value reliably
If you want delivery rather than advice I offer fixed‑scope delivery and custom builds; see Services and Custom build to compare options. Read why I take an ops‑first approach on About the author.
Where businesses waste the most time
- Re‑keying customer dataMultiple staff hours per week, frequent data errors and customer frictionCreate a direct connection so the customer record updates automatically between the systems you already pay forA person still checks: Person verifies flagged mismatches and edge cases
- Invoice chase and approvalsLate payments, manual chasing and lost interest on receivablesConfigure approval routing and automatic reminders inside your existing finance systemA person still checks: Finance approves invoices that exceed threshold amounts
- Month‑end report assemblyOvertime, delays in closing the books and single‑person dependencyAutomate the data pulls into the spreadsheet and publish results to the team dashboard you already useA person still checks: Finance reviews anomalies and signs off the final pack
How this compares with what you already run
| Operational area | The manual way | The automated way | Annual business impact |
|---|---|---|---|
| Re‑keying customer data | People copy data between systems | Data syncs automatically between systems | Lower payroll and fewer entry errors annually |
| Invoice approvals | Email chains and manual chasing | Automated routing and reminders | Faster payments and fewer late fees yearly |
| Month‑end reporting | One person assembles spreadsheets | Reports assembled and reconciled automatically | Shorter close and less overtime pay each year |
| Customer onboarding | Manual account creation and checks | Steps triggered automatically with manual sign‑off | Faster onboarding and earlier revenue recognition |
| Support triage | Agents forward and tag tickets manually | Tickets routed with suggested replies | Lower handling time and fewer escalations annually |
| Order fulfilment checks | Manual verification before shipping | Checks performed automatically with alerts | Fewer shipping mistakes and returns per year |
What the end-to-end workflow looks like
- IdentifyChoose one high‑cost, repeatable process and map the exact steps people follow today.Stays human
- Define successWrite acceptance criteria as measurable business checks, not vague promises.Stays human
- PrototypeBuild a narrow pilot that covers the happy path and common exceptions.Automates routine steps while logging exceptions
- ShadowRun the automation in parallel with the manual process to compare outcomes.Automates but keeps manual fallback
- HandoverDeliver runbooks, monitoring, and a training session for the ops owner.Automates routine runs; humans handle exceptions
- OperateMonitor performance, handle alerts, and budget scheduled work for improvements.Automates routine runs; humans handle exceptions
- IterateCollect metrics and update the automation when business rules change.Automates updated routine steps
The metrics that tell you it worked
| Metric | What it tells you |
|---|---|
| Automated hours saved per month | Sum of staff hours no longer required because of automation; shows direct payroll impact |
| Time to close month | Days to complete month‑end work; measures speed and reduction in overtime |
| Error rate on processed items | Share of items needing human correction after automation; measures reliability |
| Number of production incidents per quarter | Count of failures that require manual intervention; measures operational maturity |
| Time to acknowledge failures | Average time from an automation failing to someone beginning diagnosis; measures support responsiveness |
| Percentage of processes with documented acceptance criteria | Share of automations that have measurable business tests; measures governance |
| Cost to fix production failures | Labour cost spent per quarter fixing automation breakages; shows hidden maintenance cost |
When this is the wrong move
- When the process is low volume and irregular — automation is unlikely to pay back.
- When the process is fundamentally flawed — automating it locks in bad practice.
- When you have no one to own the automation after delivery — a pilot will die.
- When regulatory requirements demand vendor certification and you lack the budget for certification.
- When you want a platform capability but have no budget for ongoing operations.
The mistakes that cost the most
- No acceptance criteriaTeams demo a working pilot and call it done; months later it fails in production because no one defined what 'works' looked like at scale.
- No ops ownerA consultant delivers and leaves; nobody monitors alerts, so small breaks become large outages and the automation is disabled.
- Choosing tools over processLeadership buys a shiny app that covers part of the process; the remainder is glued with spreadsheets and manual steps, creating new headaches.
- Hiring prematurelyYou hire one person expecting dozens of automations; they spend their time firefighting and never build a repeatable delivery pipeline.
Frequently asked questions
When should I buy an off‑the‑shelf app instead of hiring a consultant?
Buy when the process maps exactly to the app and the app's first‑year TCO is lower than consultant delivery plus switching cost. If users can be trained in a week and configuration is minimal, buying is faster; otherwise hire a consultant for a fixed scope.
How do I know a consultant will hand over cleanly?
Require measurable acceptance criteria, a documented rollback plan, and a 30–90 day shadow period in the contract. Also ask for runbooks and at least two knowledge transfer sessions so multiple people can operate the automation.
What exact signal justifies hiring a full‑time automation person?
When monthly hours saved exceed 160 and the dollar value of those hours surpasses the fully loaded cost of the hire. Also hire when you have at least five medium‑complex automations lined up or frequent process changes needing ongoing development.
How much should I budget for maintenance after delivery?
Budget roughly 20–40% of the initial delivery cost in year one for monitoring and fixes. That range covers small tweaks and the inevitable changes when upstream systems or forms change.
Can I start with a cheap pilot and scale later?
You can, but only if the pilot has acceptance criteria, monitoring and a handover plan. Cheap pilots without those elements usually become abandoned proofs‑of‑concept that never reach production.
Stop paying salary for robotic work
Bring your messiest process to a 15-minute audit and we will map where the hours actually go, live on screen. You leave with the number whether or not you work with me.