Why start with a decision
Large cloud, data and AI programs often start from a conclusion that has not been tested: "we need to migrate", "we need a new data platform", "we need an AI assistant". The program is then scoped around that conclusion, and the first months are spent discovering whether it was right.
We prefer to start one step earlier. A two-to-four-week assessment with a fixed price lets you test the conclusion, and test the team, the method and the findings, before committing to implementation. It is a decision you buy, not a program.
What the client gets
Evidence before commitment: a view of the problem, its causes and the options, with costs and risks stated. In practice, that means a small set of concrete deliverables.
| Deliverable | What it answers |
|---|---|
| Findings register | What is wrong or at risk, with evidence, impact and an owner for each item |
| Options with trade-offs | The realistic ways forward, including doing less or nothing, with cost, risk and time |
| Recommendation | Which option we would choose and why, stated plainly |
| Roadmap | The next 90 to 180 days, sequenced, with what each step depends on |
| Executive briefing | The decision, explained in an hour to the people who make it |
If the evidence says the next step is small, or that your own team should do it, the assessment has still paid for itself. Avoiding the wrong program is usually worth far more than the assessment.
How to judge an assessment
Not every assessment is useful. Before you buy one, from us or anyone else, check four things:
- It ends in a decision. The output should make a choice easier, not describe the current state at length.
- It shows evidence. Every finding should point to something you can verify: a configuration, a bill, a test, an interview.
- It includes options you might not buy from the provider. An assessment that only ever recommends the provider's next service is a sales document.
- The people who assess are the people who would build. Otherwise the findings and the implementation are owned by different people.
What a good assessment looks like, week by week
A typical three-week assessment has a simple rhythm. The exact content depends on the question, but the shape rarely changes.
| Week | What happens | What you see |
|---|---|---|
| 1 | Kick-off, access, interviews with the people who build and run the system and with the business owner | A confirmed scope and a list of early observations |
| 2 | Hands-on review of architecture, configuration, cost data and operations; options take shape | A mid-point review of draft findings, so nothing in the final report is a surprise |
| 3 | Findings, options and roadmap written up and tested with your team | The register, the roadmap and an executive briefing |
The mid-point review is easy to skip and should not be. It is where findings get corrected, priorities get challenged and the final briefing becomes a confirmation rather than a debate.
What the assessment needs from you
- A sponsor who owns the decision and will attend the briefing.
- Read-only access to the relevant accounts, billing data and documentation in the first days.
- A few hours of the right people: the engineers who know the system, someone from finance if cost is in scope, and the business owner.
- Honesty about constraints: budget, deadlines, contracts and politics that limit the options.
When those are in place, three weeks is enough for most questions. When they are not, no amount of extra time compensates.
What it says about the provider
A provider willing to sell a focused first step, with a price and exclusions, is betting on the quality of its work rather than on the size of the contract. That is the relationship we want with clients. If the first engagement is good, the next one follows from the evidence. If it is not, the client has lost weeks, not a year.
It also keeps the provider honest about scope. A fixed-scope assessment has to state what it covers and what it does not, what the client must provide and how the result will be accepted. Those are the same disciplines that make implementation go well.
What happens next
Most assessments end with a roadmap and a recommendation for the next phase. From there, there are four common paths:
| Path | When it fits |
|---|---|
| Do nothing yet | The risk is lower than feared, or the timing is wrong |
| Your team implements | The work is clear and your team has the skills and capacity |
| We implement in milestones | The stakes are high and you want the people who assessed it to deliver it |
| Another provider implements | The work needs scale or capabilities outside our focus, such as large migrations or 24×7 operations |
When we do implement, it is priced in milestones, each with acceptance criteria, so every commitment is informed by the one before. Our article on fixed scope or time and materials explains how to structure that, whoever does the work.
Questions we hear
Isn't an assessment just a delay?
It takes two to four weeks. Starting the wrong program, or the right program in the wrong order, usually costs months. The assessment also produces the scope, acceptance criteria and plan the implementation needs anyway.
What if we already know what we want to build?
Then the assessment is short and confirms it, or it finds the one assumption that would have hurt. Either way, the implementation starts with fewer unknowns.
Do we have to use you for implementation?
No. The deliverables are yours and are written so that your team or another provider can act on them.