Test Whether You Can Leave a Software Vendor Before You Buy
Prepared with AI assistance by Alactic Editorial. Examples identified as hypothetical are illustrative, not customer results.
A vendor shows you an export button during a sales demonstration. A spreadsheet downloads, the rows look familiar, and portability gets marked as covered. But the useful question is not whether a file comes out. It is whether another team could use that file to continue the work without reopening the original application.
Before buying, ask for a representative sample export, inspect its relationships and context, and give it to someone who did not prepare it. Have that person reconstruct a real business situation from the exported materials alone. Then record what was preserved, what needed explanation, who controls the export, and which uncertainties should change the purchase decision.
This is a pre-purchase feasibility test, not a detailed migration design. You do not need to choose a replacement system or map every field. You do need enough evidence to distinguish a usable exit path from an untested assumption.
1. Define what someone must be able to continue
Start with the work, not a list of file formats. A contact spreadsheet may preserve names and phone numbers while losing the reasons those people matter, the commitments made to them, and the colleague responsible for the next action.
Write a short continuity statement: A receiving team must be able to identify active work, understand its current state, find supporting material, and assign the next action without consulting the old workspace. Adapt that statement to the purchase under consideration.
For a customer-management system, the essential package might include accounts, contacts, open opportunities, activities, notes, attachments, owners, and the relationships among them. For a project tool, it might instead centre on tasks, dependencies, decisions, and deliverables. These are suggested test categories, not claims about any vendor’s export features.
Separate essential continuity from desirable fidelity:
- Essential: Information needed to resume active work correctly.
- Useful: Historical context that reduces investigation or manual reconstruction.
- Acceptable to lose: Presentation details or convenience features that do not change the meaning of the work.
This prevents two opposite mistakes: accepting a thin export because it opens in a spreadsheet, or rejecting a workable product because its original interface cannot be reproduced elsewhere.
2. Request a representative package, not a clean demonstration
Ask the vendor to export a small, deliberately awkward set of fictional records from a demonstration or evaluation workspace. Confirm what is available in the workspace and plan being considered. A different plan, administrator role, or separately purchased service may change the proposed exit route; treat those differences as questions to resolve, not assumptions.
The sample should exercise the information your team would find difficult to reconstruct:
- An organisation with several associated people and more than one active item of work.
- A reassigned record whose earlier activity belongs to another user.
- A custom status and a custom field containing a meaningful value.
- A note with line breaks, an attachment, and an activity with a timestamp.
- A record with a missing optional value, so blanks can be distinguished from export failures.
- A closed item alongside an open one, to test whether scope filters are understood.
Request the exported files, a field explanation, relationship keys, and a description of anything intentionally excluded. If the proposed route uses an API or vendor-assisted extraction, ask to see a sample of that route rather than treating an ordinary screen export as equivalent evidence.
Use synthetic information wherever possible. The test should establish usability without creating an unnecessary copy of live customer or employee data.
3. Run a cold handoff before interpreting the results
Give the package to a colleague who did not build the sample. Include the documentation intended to accompany an eventual exit, but do not narrate the answers. A package that works only while its creator explains it has not yet demonstrated a dependable handoff.
Ask the recipient to answer concrete questions:
- Which organisation and person does each open item belong to?
- What happened most recently, and what should happen next?
- Who currently owns the work, and who authored the historical notes?
- Can the supporting attachment be opened without the original workspace?
- What do the status values and timestamps mean?
- Which conclusions cannot be drawn from the package?
Observe where the recipient stops, guesses, or asks for access to the source application. Record those moments as findings. Do not quietly fix the package during the test: preserve the original result, document the repair, and repeat the affected check.
A spreadsheet or basic file viewer can be enough for this exercise. You are testing whether meaning survives the handoff, not whether an import into an unnamed future system will be effortless.
4. Filled portability review: a fictional worked example
Everything in this worked example is hypothetical. The organisation, workspace, sample quantities, findings, and purchasing decisions are fictional. They illustrate how to record evidence rather than describe a tested product.
Purchase under review: A customer-management workspace for a small account team.
Continuity requirement: A receiving team must resume open account work, identify the responsible person, and read the latest commitment with its supporting document.
Sample: Three organisations, six contacts, four opportunities, nine activities, and two attachments.
Review roles: Operations prepares the sample; a sales coordinator performs the cold handoff; IT checks administrative access.
Finding A: The records are present, but a relationship is ambiguous
Sample check: Compare the sample inventory with the exported organisation, contact, and opportunity files. Trace each opportunity back to its organisation and associated contact.
Observed result: All sample records appear. Opportunities include stable organisation identifiers, but associated contacts are represented only by names. Two sample contacts share the same name, so the receiving colleague cannot identify the correct person confidently.
Missing context: A stable contact relationship key, or a separate relationship file.
Ownership question: Who can obtain a relationship-complete export, and is that route available under the proposed purchase?
Purchase consequence: Hold acceptance of the portability requirement. Request a revised sample with unambiguous links. If unavailable, assess whether a manually maintained relationship register is realistic; do not count name matching as a verified solution.
Finding B: Status values are readable but not understandable
Sample check: Ask the receiving colleague to identify which opportunities need action and explain a custom status value.
Observed result: The status column contains codes. The files do not explain which code means awaiting customer confirmation. The workspace administrator can explain it verbally.
Missing context: A dated status dictionary, including the meaning of custom values and the distinction between active and closed work.
Ownership question: Will the vendor provide this dictionary, or must the internal process owner maintain it?
Purchase consequence: Proceed conditionally if an internal owner accepts the documentation task and a repeat handoff succeeds using the dictionary alone. Treat the ongoing documentation effort as part of the operating cost.
Finding C: Attachment links do not constitute an independent package
Sample check: Open both exported attachment references without an authenticated session in the source workspace.
Observed result: The references lead back to the vendor workspace and require access there. No attachment files are included in the sample package.
Missing context: A demonstrated way to retrieve the files and preserve their association with the relevant records.
Ownership question: Can an authorised administrator obtain the files in bulk, or does the proposed route require vendor assistance?
Purchase consequence: Do not approve the exit assumption for supporting documents. Request a file-inclusive sample. If that cannot be demonstrated, either keep essential documents in a separately controlled repository and test the links, or reject the product for this workflow.
Finding D: Historical authorship becomes misleading
Sample check: Compare a reassigned opportunity’s owner with the authors of its earlier notes.
Observed result: The current owner is named, but historical activities contain unexplained user identifiers. The recipient initially attributes an earlier commitment to the current owner.
Missing context: A user lookup that preserves historical authorship, including users no longer active in the workspace.
Ownership question: Who can supply the lookup, and what happens to exported references after a user is deactivated?
Purchase consequence: Require a corrected sample or a tested internal lookup process before approval. If historical authorship is essential and neither works, the gap is disqualifying.
Finding E: Timestamp interpretation remains uncertain
Sample check: Put the sample activities in order and identify the latest customer commitment.
Observed result: Dates and times are present, but the package does not specify the time zone. The sequence is understandable in this sample; exact timing is not.
Missing context: Timestamp definitions, including time zone and the meaning of created, updated, and activity dates.
Ownership question: Who will obtain and preserve the field definitions?
Purchase consequence: Keep timing-dependent use cases unresolved until definitions arrive. If exact timing is not essential to this workflow, document the limitation explicitly rather than calling the whole export complete.
Finding F: Only the evaluation owner has demonstrated access
Sample check: Ask a second authorised administrator to repeat the documented export procedure.
Observed result: The second administrator cannot access the demonstrated route. It is unclear whether this reflects role configuration, plan availability, or a special evaluation arrangement.
Missing context: Confirmed production access requirements and a repeatable procedure.
Ownership question: Which internal role will own export access, who is the backup, and when is vendor involvement required?
Purchase consequence: Pause approval until the access route is clarified and demonstrated under the proposed configuration. A salesperson’s successful demonstration does not resolve internal ownership.
5. Make ownership part of the test
An export path is operationally weak if nobody knows who can use it. Assign an internal owner and a backup, then ask each to explain the procedure and its dependencies. Document whether it needs workspace access, a particular role, vendor assistance, or a separately managed file store.
Do not solve this by giving everyone broad administrative privileges. OWASP’s authorization guidance recommends least privilege. The practical objective is to give designated people the permissions they need, not to make unrestricted access a condition of continuity. This is design guidance, not evidence about a particular vendor.
Also ask who maintains the supporting context: status definitions, user lookups, field meanings, and export instructions. If those materials live only in one person’s memory, the exit package remains incomplete even when the raw records are available.
6. Record uncertainty without turning it into a pass
Use four evidence states in the purchase review:
- Demonstrated: The receiving colleague completed the defined check from the supplied package.
- Conditional: A specific repair or internal workaround is proposed, with an owner and a required retest.
- Unknown: The route has not been demonstrated or the answer remains ambiguous.
- Unacceptable: A necessary continuity requirement cannot be met by the proposed approach.
Keep statements of capability separate from observed results. A vendor’s statement that files can be supplied is useful follow-up information; it is not the same evidence as receiving and opening a representative package.
A small successful sample also leaves scale uncertain. Ask how a larger extraction would be performed, what assistance it requires, and what constraints remain untested. Do not extrapolate a completion time from a tiny demonstration or demand a full migration rehearsal for a preliminary purchase decision.
7. Choose between repair, workaround, and rejection
Not every gap should stop a purchase. A missing status dictionary may be cheap to maintain internally. Missing relationships across active work may require substantial reconstruction. The distinction is the burden and consequence of the gap, not whether the vendor calls it an export limitation.
Evaluate workarounds for their own failure modes. A separate document repository can reduce dependence on exported attachments, but it creates a linking discipline that someone must maintain. A manually updated user lookup may work for a stable team but become unreliable if ownership changes frequently. Test the workaround rather than merely naming it.
When a check fails, preserve the failed sample, identify the smallest corrective change, and repeat that check with the receiving colleague. If the repair requires ongoing manual work, name its owner and decide whether the team will actually sustain it.
Keep this review bounded. It does not establish legal rights, define retention policy, or guarantee compatibility with every future replacement. It establishes what your team has evidence it can recover and understand.
8. Make the next purchasing step measurable
Before the next vendor review, create a portability decision record with a continuity statement, sample inventory, findings, evidence states, named owners, and purchase consequences. Book the cold handoff before marking portability complete.
For the fictional review above, the decision is not ready for approval: contact relationships, independent attachment access, and administrator access remain unresolved. The status dictionary is a conditional internal task, while timestamp meaning still needs clarification.
Your next action is specific: send the unresolved checks to the vendor, request the missing sample materials, and schedule a repeat handoff. Record how many essential checks the receiving colleague completes without help, which require source-workspace access, and which still require guessing. That gives the buyer an observable basis for proceeding, accepting a defined limitation, or walking away.
Put the next conversation to work.
Explore Alactic Communications for WhatsApp conversations, campaigns and AI-assisted workflows. Choose the product and plan that fit your team.
Explore Communications