How to Scope a Spreadsheet Automation Project

The quality of a quote depends on the quality of the brief. A vague request such as "automate our monthly report" leaves the builder guessing, and guesses become either padding in the price or surprises halfway through. An hour or two of preparation on your side produces a clearer quote, fewer change requests and a result that fits how your team actually works. Here is what to prepare, step by step.

Step 1: Write down the current workflow, step by step

Start with what happens today, not what you want the tool to do. Walk through one real cycle, ideally while someone performs it, and record every step, including the small ones people do from habit.

A simple table works well:

#	Step	Who	File or system	Input	Output	Time
1	Export open invoices	AR clerk	ERP	Date filter	invoices.csv
2	Paste into tracker	AR clerk	Tracker.xlsx	invoices.csv	Updated sheet
3	Fix customer names	AR clerk	Tracker.xlsx	Manual lookup	Clean names
4	Build aging summary	Analyst	Pivot table	Tracker	Summary tab
5	Copy charts to slides	Analyst	PowerPoint	Summary tab	Monthly deck

Fill in the time column with your own estimates. Watch for steps such as "check it looks right" or "fix the ones that don't match." They usually hide the business rules that matter most, and they need to be spelled out for an automated version to reproduce them.

Step 2: Collect representative sample files

Samples are the most useful thing you can provide. Choose files that show the real range of what the process handles:

• A typical period.
• A difficult period: year-end, a month with corrections, or the largest file you have had.
• The current output, such as the finished report or deck, so the builder can see the target.
• Any lookup or mapping tables people maintain by hand.

Remove confidential data without breaking the structure

Replace names, account numbers and personal details with consistent placeholders: "Customer 001" should stay "Customer 001" in every file, so joins still work. Keep column names, sheet names, formats and the quirks of the data (blank rows, merged headers, text-formatted numbers), because those quirks are exactly what the automation must handle. If amounts are sensitive, scale them by a constant factor rather than replacing them with random numbers, so totals still reconcile.

Before sending, run File > Info > Check for Issues > Inspect Document in Excel to find hidden sheets, comments, document properties and personal information you may have missed.

Step 3: Define the output and what "done" means

Describe the finished result precisely: file type, layout, which tabs or slides, who receives it, and where it is saved. Then write acceptance criteria you can check on the day of delivery. Good criteria are specific and testable, for example:

• Totals on the summary tab match the source system's report for the two sample months.
• The process runs from one button or one refresh, with no manual pasting.
• New customers that are not in the mapping table appear on an exceptions tab instead of being dropped.
• A team member who did not build it can run it by following a one-page guide.

"Faster" or "easier" are goals, not acceptance criteria. If you cannot test it, it will be argued about later.

Step 4: List exceptions and edge cases

Most automation effort goes into the cases that happen occasionally. Ask the people who do the work what goes wrong, and list each case with what should happen:

• A source file is missing or arrives late.
• A column is renamed or moved after a system update.
• Credit notes or negative values appear.
• The same transaction is exported twice.
• A period has an extra week, or a fiscal calendar differs from the calendar month.
• A new product, region or cost center appears that is not in any lookup.

For each, decide whether the tool should handle it automatically, flag it for a person, or stop with a clear message.

Step 5: Record your environment

The same solution can be simple or complicated depending on where it runs. Note:

• Excel and Office versions: Microsoft 365 or a perpetual version such as 2019 or 2021, and Windows or Mac. Some functions and Power Query features differ between versions, and VBA behaves differently on Mac.
• Macro policy: whether IT allows macro-enabled workbooks, and whether they must be signed.
• Where files live: a local drive, a network share, OneDrive, SharePoint, Teams, or email attachments.
• Other tools available: Power BI, Power Automate, or Python, and whether your licenses and IT policies allow them.
• Who runs it: one person, a team, or an unattended schedule.

Step 6: Decide what stays manual

Not every step should be automated. Approvals, judgment calls, and exceptions that happen once a year are often cheaper and safer to leave with a person. Mark each step in your workflow table as "automate," "assist" (the tool prepares, a person confirms) or "keep manual." This single column has a large effect on both scope and price.

Step 7: Estimate frequency and time spent

Note how often the process runs (daily, weekly, monthly, ad hoc), how long each cycle takes, and how many people touch it. Use your own numbers, even rough ones. They help you judge whether a given quote makes sense, and they help the builder decide where to focus, for example on the step that takes most of the time or causes most of the errors.

What a good fixed quote should contain

Once you share your brief, a fixed quote should make clear exactly what you are buying. Look for:

• Deliverables: each file, report, macro, query or flow, named specifically.
• Assumptions: what the price is based on, such as file structure, versions, volumes, and which edge cases are included.
• Exclusions: what is not covered.
• Acceptance criteria: how you and the builder will agree the work is done, ideally matching what you wrote in Step 3.
• Timeline and milestones: when you see a working draft and when final delivery happens.
• Change process: what happens if requirements change after the quote is accepted.
• What is needed from you: sample files, access, review time.

If a quote lists only a price and a one-line description, ask for the assumptions in writing before you start.

How milestone payments protect both sides: a worked example

Here is how the process works at 365 Metrix, as one concrete example.

1. You describe the task in the Discovery Studio. An AI analyst asks follow-up questions and drafts a specification, which you review and correct.
2. A human specialist reviews the specification and your sample files, then sends a fixed quote that lists the deliverables and assumptions. The price is fixed, never billed by the hour.
3. Payment is split into three milestones: 30% at project start, 40% when you approve the working draft, and 30% at final delivery.

The split protects both parties. You are not paying the full amount before you have seen anything work, and the largest payment depends on your approval of a working draft built on your own sample files. The specialist, in turn, has a commitment before work begins and a clear checkpoint where scope is confirmed, so late changes can be discussed openly instead of absorbed silently.

Uploaded files are stored privately and are never opened or run on the server. If you want a rough idea of cost before writing a full brief, the project estimator gives a preliminary, non-binding range with every rule shown, and the pricing page explains how fixed quotes are structured.

Limitations and trade-offs of fixed-price scoping

• Scoping takes your time. The steps above can take a few hours, and they require access to the people who do the work.
• A fixed price carries risk on both sides. The builder has to price in some uncertainty. A clearer brief reduces that uncertainty.
• Assumptions can turn out wrong. If the samples are not representative, the quote may need revisiting once real data arrives. That is why the difficult period matters in Step 2.
• Changes cost something. Fixed scope means new requirements become a change request, not a free addition. That is the price of predictability.
• Not every process should be automated. If a task runs twice a year and takes an hour, the honest answer may be a checklist, not a tool.

If your process spans Excel, email and shared folders rather than a single workbook, see workflow automation services.

When to bring in a specialist

If you have mapped the workflow and still cannot tell which parts are worth automating, or the process spans several tools and owners, a specialist can help draw the boundaries. When you are ready, describe your project and attach your sample files; you will receive a specification to review and then a fixed quote.