Three months after we left, the model stopped working
During a cost transformation, your team built a detailed cost model in Excel: dozens of tabs, lookup formulas, macros, and a monthly data pull from the client's finance system that one analyst did by hand. At handover, the client's finance team received the file and a walkthrough. Three months later, the client calls: the numbers look wrong and nobody knows why. A column in the finance export changed, and the model silently picked up the wrong figures.
The client is unhappy, and your firm either fixes it for free or leaves a poor impression.
Why handed-over models fail
- Models are built quickly under engagement pressure, not for long-term use.
- Data pulls are manual steps that live in a consultant's head.
- Formulas are complex, undocumented and fragile to changes in source data.
- Checks that would reveal broken inputs do not exist.
- Client teams receive a walkthrough but not a manual or training.
Commercial pressure pushes the problem to the end. Budgets rarely include time for building a model properly, so it is built quickly to get the answers, and handover is squeezed into the final week alongside everything else.
What fragile handovers cost
| Issue | Effect |
|---|---|
| Model breaks silently | Client makes decisions on wrong numbers |
| Client asks for help | Unbilled support or awkward conversation |
| Model abandoned | Benefits of the engagement fade |
| Poor impression | Follow-on work less likely |
The reputational cost can outlast the engagement. The client's finance director does not remember the elegant analysis; they remember the model that gave the wrong answer at a budget meeting. Future proposals to that client start from that memory.
How we turn engagement models into tools
- We review the model with your team to understand inputs, logic and outputs, and agree what needs to be kept.
- Manual data pulls are replaced by connections to the client's systems, or by structured import templates that validate each file.
- Logic is simplified where possible and moved into a clear structure, in Excel, Power BI, or a small web application, depending on who will use it.
- Checks are added that flag missing, changed or out-of-range data before results are produced.
- Documentation covers how to run it, what each input means, and what to do when a check fails.
- The client team is trained, with your consultants, on running it themselves.
- Optionally, a support arrangement is set up between your firm and the client, if you want to offer one.
We work as a technical partner to your firm, behind the scenes or alongside your team, as you prefer. The client relationship stays yours.
We usually look at one or two recent models that clients have struggled with, to understand where they break, before agreeing a standard approach your teams can use on future engagements. For live engagements, we can start as soon as the model's structure has settled.
What the handover looks like afterwards
The client receives a tool that runs, with checks that tell them when something is wrong, and a team trained to use it. When the source data changes, the check fails loudly instead of the numbers quietly going wrong.
Your firm's work keeps delivering value after the engagement, which is what leads to the next one.
Does this sound familiar?
- Clients struggle to run models after handover.
- Models depend on manual data pulls.
- Handover means a walkthrough and a file.
- You have fixed a client's model for free after an engagement.
- Nobody knows if handed-over tools are still used.