# Turning a shipped deployment into a case study: agree the publishable numbers and the sign-off before you write

> A customer champion who says yes to a case study is not the same as their company giving permission, and your best number may be the first one struck out.

Original: https://fdetimes.net/en/guides/deployment-case-study-approval-metrics/

The project has just settled into stable production. The customer's champion is happy and says "go ahead and write the case study". Three weeks later the draft comes back from their legal team with two of the three headline numbers struck out, and the story you built around those two numbers collapses with them.

Two different things are easily mistaken for one here. The first is being liked by the users. The second is being allowed by their company to say so in public.

As an FDE, you are usually the person who knows each number in a deployment best: how it was measured, on what data, over what period. That is why asking for permission to publish should be treated as part of the engagement, as important as measuring the results, rather than something left until after handover.

## Who actually has the authority to say yes?

Pravin Kumar, writing about how to get customer approval for SaaS case studies, points out something easily overlooked: the person who loves your product is not the person who approves public statements. In his account, final sign-off comes from legal, PR or a senior executive.

Manifestly's case study production checklist is blunter still. A champion's verbal agreement is not permission, because procurement, communications or legal often block publication at a later stage.

So the first question to ask is not "do you agree?" but "on your side, who approves public content that names the company?".

The answer also shapes the schedule. The Content Marketing Institute notes that customers may already have had to consult legal or leadership just to get permission to take part. If you plan to publish two weeks after finishing the draft, you are betting on a process you do not control.

## Which numbers survive review?

Logo and metrics need separate approval. Manifestly records customers who allowed their logo to be used but refused to publish hard numbers. Kumar adds that some companies will share percentages but not absolute figures. Manifestly's checklist advises agreeing each form in advance: percentage or raw figure, range or exact value, or an anonymised industry benchmark.

Consider a hypothetical example. You put an invoice-reading agent into production for a distribution company. After three months, processing time per invoice has fallen from 12 minutes to 3, the system handles 40,000 invoices a month, and 9 of the 15 people on the data-entry team have been moved to other work.

Each of these numbers carries a different level of risk for the customer. "Down from 12 minutes to 3" can be written as "a 75% reduction" and keep all its weight.

"40,000 invoices a month" reveals the scale of the business, which the customer's competitors would very much like to know. The staffing change is the most sensitive area, because it touches internal communications and relations with employees.

| Internal metric | Form to propose for publication | Who to consider asking (varies by customer) |
|---|---|---|
| 12 minutes down to 3 per invoice | "75% reduction in processing time" | Champion, communications |
| 40,000 invoices/month | "Tens of thousands of invoices a month", or drop it | Leadership, legal |
| 9 of 15 people reassigned | Talk only about team time freed up, or drop it | PR, senior leadership |
| Company name and logo | Anonymised by industry if not approved | Legal, marketing |

This table cannot predict what the customer will approve, and the last column is only a starting point for questions to the champion, not a list of the real approvers. What it does is give you a fallback ready for every number, so that if one row is struck out the piece still keeps its main point.

**Key point:** Build the story around the number the customer will definitely allow, not the most impressive one.

## Where does the FDE sit in the approval loop?

Three groups take part in a case study. Marketing writes and publishes it. The AE or CSM owns the customer relationship, and Manifestly advises that the draft go through them rather than arrive from marketing as an email from a stranger. The missing piece is someone accountable for the numbers, and that is the most natural place for the FDE.

There is a clear reason for this division of labour. When the customer's legal team asks "how was the 75% reduction calculated?", marketing or the AE usually cannot answer. The person who wrote the queries and built the dashboard can answer at once. Take on this role proactively; do not wait to be asked.

## The process from project to signed-off draft

The first step should happen as soon as results are measured: draw up a list of metrics like the table above, noting next to each one its data source and the proposed form for publication. Later, those data sources become the substantiation file.

Second, ask the champion to identify the approver, ask how long approval takes and build that time into the plan. Third, following Kumar's advice, agree in writing what can and cannot be shared before you start writing.

A single email summarising the metrics table, answered "agreed" by someone with authority, is worth far more than a yes over the phone.

Fourth, write the draft using only the agreed forms of each metric, and send it with a substantiation file containing the queries, dashboards and measurement periods.

Manifestly advises that this file always travel with the document because in the US there is legal risk around advertising under the FTC and the Lanham Act. Even if you work in another market, the habit lets you answer any "where does this number come from?" question quickly.

Fifth, pass the draft to the AE or CSM who owns the relationship so they can send it to the customer. Finally, as FormAssembly advises, send the finished version back for the customer to read before publication, because their marketing or legal team may need to approve further details.

## Mistakes that leave case studies abandoned

The most common mistake is treating the champion's word as a licence, then discovering in the final week who really signs off. The second is asking about the metrics only after the draft is finished: when the central number is cut, you have to rewrite from scratch instead of changing one line.

The third is bundling logo and metrics into a single approval request, so the customer rejects the whole package over one detail. The fourth is keeping no substantiation file, so when the customer's legal team asks how the "75% reduction" was calculated, nobody can answer the same day.

## How this skill shows up on a CV

The results lines on your CV and in your interviews should follow the same discipline. They need to show that you can measure value, without turning you into someone who leaks a former customer's data.

"Cut invoice processing time by 75% for a distributor" is both persuasive and safe. Stating a former customer's transaction volume does not make the story stronger; it only adds risk.

In an interview, if you are asked how you prove value to customers, walk through the approval process above, with an example of a number you converted into a safe form. That answer shows the interviewer that you know how to measure and how to keep a customer's trust.

This week's exercise: take the most recent project you shipped, build a table like the one above and check whether the last column contains real names yet. If it does not, that is the job to do before anyone starts writing.

**Try this week:**

- Pick a project you shipped, list five result metrics and note beside each the safest form in which it could be published (percentage, range or anonymised).
- Ask the AE or CSM who owns that customer: who on the customer side approves public statements, and how long approval usually takes.
- Rewrite one results line on your CV using only a form of the metric that reveals nothing internal about the customer.

## Sources

- [How to Get Customers to Participate in a Case Study](https://contentmarketinginstitute.com/articles/customers-participate-case-study/)

- [How Do You Get a Customer to Approve a SaaS Case Study? | Pravin Kumar](https://www.pravinkumar.co/blog/how-to-get-a-customer-to-approve-a-saas-case-study-2026)

- [Workflow Templates for Marketing: Case Study Production Checklist (Manifestly)](https://app.manifest.ly/public/checklists/case-study-checklist)

- [A 6-Step Process for Creating an Effective B2B Case Study](https://www.formassembly.com/?p=12974)