An SME roadmap often arrives as a polished deck: three horizons, a maturity curve and a list of platforms. Six months later the same manual workarounds survive, because the deck described destinations without naming the constraint that had to move first.
This buying guide is about purchasing the work that produces a usable sequence: a diagnosis, a set of costed options, a stop list and a governance routine that survives the next budget cycle.
Define the decision before the deliverable
The board usually needs to decide four things: which constraint to relieve first, what to stop funding, what must be true before the next stage, and who owns each commitment. Write those questions into the brief. A roadmap that cannot answer them is a communications exercise.
Set a horizon you can govern. Twelve to eighteen months is long enough to sequence a migration and short enough to keep owners accountable. Anything longer becomes a wish list with dates attached.
What a credible roadmap contains
- A baseline. Current cycle times, error rates, cost lines and customer measures, each with a named source.
- A process inventory. Where work crosses teams, systems or paper, and where it waits.
- A systems and data map. What holds the customer, price, stock, invoice and consent record.
- A constraint analysis. The one or two limits that make everything else slower.
- Options. At least three, each with an effort model and a total cost of ownership.
- Sequencing rules. Why this before that, and what triggers the next phase.
- A capability plan. Who will run the new process, what training is needed and what stays outsourced.
- Governance. Decision rights, review cadence and the conditions under which the plan stops.
Five ways to buy the work
Strategy consultancies are strong at diagnosis and board communication, and weaker at implementation detail unless the same team stays for delivery. Ask who will be in the room during rollout.
ERP or software vendors know their own product deeply and will naturally see the roadmap through that lens. That is useful when the platform decision is already made, and risky as the only voice in the room.
Systems integrators are good at connecting existing systems and estimating integration effort. Check whether the reference work matches your scale, and whether the team that sells is the team that delivers.
Boutique automation or software firms go deep in one value stream and are often more candid about what will not work. They are less useful when the problem spans finance, operations and customer channels at once.
An internal team with a facilitator has the best context and the lowest cost to sustain, provided a senior person can protect the time. Bring in an external reviewer to challenge the diagnosis rather than write it.
Scoring the roadmap proposal
| Criterion | Evidence to ask for | Weight |
|---|---|---|
| Diagnosis quality | Baseline data, interviews and process evidence | High |
| Sequencing logic | Explicit rules, triggers and dependencies | High |
| Cost realism | Effort model, run cost and internal time | High |
| Vendor neutrality | No hidden reseller margin or single-product agenda | High |
| Capability transfer | Named owner, training and documentation | Medium |
| Risk and security | Data handling, access and continuity plan | Medium |
| Change management | Adoption plan, not a communication slide | Medium |
| Milestone definition | Acceptance criteria for each phase | High |
Weight the criteria before proposals arrive. Ask each bidder to answer the same eight questions in two pages; the quality of that answer predicts the engagement better than the pitch.
Warning signs in a proposal
- A technology list before a problem statement.
- No baseline, or numbers with no source.
- Benefits expressed as percentages with no calculation behind them.
- No run cost after the first year.
- No stop list, because everything is a priority.
- One vendor product at the centre of every option.
- Owners named as departments rather than people.
- No acceptance criteria for the diagnostic itself.
Buy the diagnostic first
A six-week diagnostic is a lower-risk purchase than a full roadmap, and it tells you whether the provider can think. Weeks 1–2 cover interviews and data collection. Week 3 maps processes and systems. Week 4 shapes options around the constraint. Week 5 builds the effort and total-cost model. Week 6 reviews the sequence with the people who will own it.
Acceptance should be written before work starts: finance and operations agree the baseline, three options are costed with the same structure, the first ninety days have named owners, and the stop list is signed. If the diagnostic produces only slides, do not buy phase two.
Contract terms that protect the sequence
- Milestone payments tied to accepted outputs, not to elapsed time.
- Ownership of working files, models and interview notes.
- Access to the people who did the analysis, not only the partner who sold it.
- Change control for new business units or additional analysis.
- Subcontractor approval and confidentiality terms.
- Exit support, including a handover pack and a final review meeting.
Sequence the first year by constraint
Most SME sequences work better in this order: relieve the constraint that limits throughput, clean the data the next system depends on, integrate one pair of systems at a time, buy a product before commissioning a custom build, and measure for ninety days before expanding.
Place the stop list beside the plan: the report nobody reads, the tool nobody logs into, the approval that adds a week without reducing risk. A roadmap that only adds work will fail on capacity, not on technology.
Payback logic you can defend
Payback months = one-off cost ÷ monthly net benefit
Net benefit is the sum of measurable time released, error and rework reduction, avoided cost and any revenue lift you can attribute. Use your own baseline rather than a vendor benchmark. If the benefit cannot be measured after ninety days, treat it as a hypothesis, not a saving.
Metrics to review quarterly
Cycle time, error rate, cost per transaction, adoption by role, data quality, milestone hit rate and benefits realised against plan. Add one customer measure so the programme is not judged only by internal efficiency.
Where this fits in our work
Roadmap work that ends in a system sits within Custom Software & Integrations, with custom web applications covering the build phase. The pricing page explains how discovery and delivery are scoped. For the four roadmap approaches this guide helps you choose between, read the comparison of SME transformation roadmaps.
Sources
- OECD, SME digitalisation, accessed 17 November 2026.
- World Bank, Digital development, accessed 17 November 2026.
- HM Treasury, The Green Book: appraisal and evaluation in central government, accessed 17 November 2026.
- NIST, Cybersecurity Framework 2.0, accessed 17 November 2026.
Responses (0)
No comments yet. Add the first one below.
Leave a response
Your email address will not be published. Required fields are marked *.