# How to run a software trial

Test one real task in every product. Record the evidence and decide whether to buy.

Canonical page: https://quillwell.com/guides/software-trial/
Updated 2026-10-29
Publisher: Quillwell Guides (https://quillwell.com/guides/about/). Editorial standards: https://quillwell.com/guides/editorial/

## Write a one-page trial brief

Write a one-page trial brief, then use it with every candidate. Name one recurring task and the people involved. Write down why your current approach falls short. “Improve reporting” is too broad. “Prepare the weekly paid-media report from three ad accounts without copying data by hand” gives you a task to observe.

Separate requirements from preferences. A requirement can be keeping client data separate or requiring approval before publication. A preference can be a particular chart style. Keeping these apart prevents an appealing interface from compensating for a missing essential control.

Write down the fallback if you buy nothing. An existing spreadsheet and a better checklist can be a reasonable option. Your trial should compare the new process with that alternative, including maintenance effort.

## Use one task across every product

Give each product the same starting material so you can compare the results. Use a representative sample you are permitted to share, and avoid uploading confidential customer data before your team has reviewed access and data handling.

1. Describe the starting state and the result you need.
2. Complete the normal task with an ordinary user account.
3. Ask a colleague to repeat it using your notes.
4. Change one input and see whether the result updates correctly.
5. Test the review or export that finishes the job. Check the handoff too.

Record where you needed vendor help. Assistance is useful, but your team might need more effort to run a workflow demonstrated by a specialist.

## Keep evidence alongside each requirement

Record what worked in your trial. Separate vendor documentation from unanswered questions. A completed task using your own inputs tells you more about fit than a sales claim alone.

| Requirement | Evidence to keep | Decision |
| --- | --- | --- |
| Refresh a weekly report | Result and time; failed-refresh behavior | Can the report owner maintain it? |
| Approve changes | Roles and a test approval | Can an ordinary user bypass review? |
| Export work | Actual exported file | Is it usable outside the product? |
| Predict the bill | Written quote and usage assumptions | What changes the total? |

Note the product and plan, then record the date and conditions of your test. Those details will help when someone asks why you chose it months later.

## Test the awkward parts before the deadline

Remove a user and reconnect an account. Correct an input, then export your work. These are ordinary events in a real team, even when they are absent from a polished demo.

For software that can modify live campaigns, use a safe test environment where available. Ask to see the boundary between a draft and a published change. Check the required permission and recovery process. Do not assume an “undo” button reverses effects that have already happened.

Ask how support handles the failure most likely to interrupt your work. Keep the answer with your evaluation, including which plan it applies to.

## Close the trial with a short decision record

At the end, decide to buy or stop. Extend the trial only to answer a specific question. Write down the tradeoffs and who will maintain the tool.

An extension needs a defined test and a deadline. “We need more time to explore” usually means the decision criteria are still unclear. For example, allow three days to validate a missing data field with support.

### Use the trial worksheet

Keep requirements and evidence together with remaining questions.

[Download the worksheet](https://quillwell.com/guides/assets/reporting-evaluation.csv)

## About this guide

Keep your requirements and trial results with the questions you still need answered.

Quillwell also builds Adswell, a paid-media product. [Read our editorial standards](https://quillwell.com/guides/editorial/).

Something missing or out of date? [Suggest a correction](https://quillwell.com/guides/request/?type=correction&page=/guides/software-trial/).