Should I charge for discovery before a fixed-price project?

Offer paid discovery when a specific unresolved question prevents a reliable project scope and the investigation will produce something useful for the customer's decision. Define the investigation as its own job, with a price, inputs, limits and deliverable. Do not sell an undefined conversation merely because the eventual project is hard to quote.

If the necessary information already exists and can be reviewed within ordinary quoting work, a separate fee may add little value. Explain what remains unknown and why resolving it requires a bounded piece of paid work.

Name the unknown that changes the quote

Consider a fictional data studio asked to migrate a retailer's 4,000 product records into a new catalog. The customer wants a fixed implementation price but has supplied only one sample file. The studio does not yet know how many source formats exist, which fields are required or what cleanup would be needed.

Quoting a reliable whole-job price requires a defined result and some basis for the required work. A low fixed amount would not make unseen data consistent. An open-ended hourly arrangement is another possible offer, but its spending uncertainty needs to be understood, as the price-basis comparison explains.

The studio proposes a separate $600 assessment, not a commitment to migrate the catalog. The customer can take the report and decline further work. No implementation order, automatic next phase or guaranteed final price is included.

Sell a deliverable the customer can use

The assessment will catalog the supplied files and formats, map required target fields, identify observed exceptions and describe the next decisions needed before an implementation quote. The customer supplies the complete source-file list and approved, appropriately redacted samples through the agreed secure route. The studio requests no production credentials or unrestricted personal data.

For this fictional job, the customer supplies four formats and file metadata totaling 4,000 records. The assessment reads 25 customer-selected sample records from each format: 100 in all. It does not inspect the other 3,900 records or claim a statistical sample representative of the whole catalog.

The output is a readable report and mapping file listing coverage, findings, exclusions and unresolved issues. Acceptance means those specified outputs cover the supplied four formats and the agreed 100 records, with the outstanding questions visible. It does not require a favorable recommendation or a successful production migration.

Budget the assessment as paid work

The complete assessment needs eight paid owner hours at $35, including intake, inspection, mapping, reporting and discussion: $280. A specialist supplies two hours of security and access review at $80, or $160. A $60 temporary tool license brings all costs to $500.

Both people have confirmed those hours within the two-day assessment window without displacing accepted work. No unpaid implementation, selling or administrative work is left outside the ten hours in this limited job. The studio has $700 separately available for the $500 payments before collecting its $600 fee at completion. No other obligations use that cash.

The price leaves $100 after those specified costs, before financing and income tax. If the report leads to no larger project, the work and fee remain the same under these supplied terms. Any future implementation has its own costs and price; the assessment's surplus is not counted again as implementation revenue.

Report what the sample actually established

Suppose twelve of the 100 reviewed records lack a required date. The report lists those twelve and the affected formats. It does not multiply twelve percent by 4,000 and declare that 480 records need repair. The selected sample supplies no demonstrated population error rate.

One finding may still change the scope immediately: who will supply missing dates, where they can be verified and whether records without them can be migrated at all. The report presents those choices before a fixed implementation promise is made.

The UK government's service-discovery guidance separates understanding the problem and constraints from committing to build a service. The studio's two-day assessment is an original commercial example, not a claim to follow that government's process or typical duration.

Leave the next decision open

The customer has a $3,000 total project allowance. Spending $600 on discovery leaves $2,400 for anything that follows; it has not bought a $3,000 implementation as well. The studio has not established that a complete migration will fit $2,400.

After the report, the customer might clean the data first, reduce the initial catalog, commission a fuller examination or stop. Any changed output and price need a fresh proposal. A useful diagnostic can justify stopping rather than become a sales device that makes the next purchase feel inevitable.

The value-of-information guide examines how information can change a choice. This assessment supplies no probabilities or guaranteed saving. Its practical test is whether the deliverable answers the promised questions and makes the next decision better informed.