- Test on a real inspection with poor signal, not in a demo or on office Wi-Fi.
- Time the whole path from the first finding to the report you would actually send.
- Include the awkward cases: a two-photo finding, a measurement with units and a client template.
- Count the taps and the duplicate entries, because that is where a trial is won or lost.
- An app that passes on a perfect day and fails on a wet one has not passed.
The short answer
What to test before choosing inspection report software is one real inspection, not a demo. Pick a job with poor signal, gloves, direct sunlight, a finding that needs two photos, and a client template you must match. Run the whole path from first finding to customer-ready report, and time it.
Sales demos run on good Wi-Fi, with clean hands, on a sample checklist that already works. Your site does not. Everything that decides whether an app survives real work shows up in that gap.
- One real job you know well, with a client waiting for the report.
- Normal conditions: the weather, access and pressure you usually get.
- The device you will carry, on the network you will have.
- The full path once, from first finding to the export you would send.
- A stopwatch, so you compare measured time rather than impressions.
- A running note of every point where you had to repeat yourself.
Set up the test properly
Decide what finished means first. Sending the report and having a draft in the app are different end points, and timing to the wrong one flatters every tool. If your end point is a file ready before you leave the site, finishing the report on site lists what has to be true.
Bring four things: the checklist you would normally use, the report order your client expects, a stopwatch, and somewhere to log friction as it happens. If your client has a fixed template, keep it open on a second device so you compare the export against it immediately rather than from memory.
Run the same job in each candidate back to back, and swap which app goes first on a second visit, because the second run always feels faster.
What breaks on a real site
These conditions separate a working tool from a convincing one. Test each deliberately. The photo row is the one most tools fail, and keeping photos linked to the finding they prove shows what a working link looks like.
- Poor or absent signal: results and photos must save with no connection.
- Gloves and cold hands: core capture actions should work one-handed.
- Direct sunlight: screen contrast decides whether you can read a checklist at midday.
- A finding needing two photos: context and defect, both attached to the same item.
- A measurement with units: free text invites an argument about what the number meant.
- A client template to match: check the exported order and headings, not just the content.
The test script, step by step
Start the clock at the first finding, not at the gate. You are timing the reporting path, not the inspection.
Capture one finding completely: a pump coupling recorded as Fail, a note, a measurement in units, and two photos linked to that item, plus your recommendation. Find it in the report preview. Turn off mobile data and Wi-Fi for ten minutes and keep working. Those ten minutes with the radio off are the part most trials skip, and what an offline inspection app has to keep on the device explains what should still be there afterwards. At the end, add the signature and produce the export your client receives, stopping the clock when that file is ready to send.
Log the numbers for each app in the table below while the site is fresh in your head.
| What to record | How to capture it | A fail looks like |
|---|---|---|
| Time to finished report | Stopwatch from first finding to the export being ready to send | The steps blur and you cannot say where the time went |
| Taps per finding | Count screen changes for one finding with a note and two photos | The defect needs a separate note, photo set or second device |
| Duplicate entry | List anything typed twice, on site and again in the report | You copy field notes into a document before sending |
| Offline behaviour | Disable data and Wi-Fi for ten minutes mid-inspection | It will not save, syncs late, or the photo arrives detached |
| Photo-to-finding link | Attach two photos to one finding and locate them in the export | Images land in a gallery at the end, unrelated to the item |
| Client template match | Export in the order and headings your client requires | Sections need reordering and reformatting before you send |
Reading what you recorded
Nobody can give you a target number of minutes, because jobs and inspectors differ. Compare the candidates against each other on the same job, and against your current workflow.
Weight failures by how often they happen. A slow export costs minutes once per report. A photo in the wrong place costs a correction and a client conversation. Duplicate entry compounds on every visit, so treat it as the most expensive habit.
The strongest signal is not the fastest run. It is the app you stopped noticing while you worked. The buying comparison sets out how the categories differ, and the side-by-side field visit method covers running two apps on one job.
How Inspection Report handles this
You can run this test in Inspection Report on the free tier: a limited inspection allowance, three starter templates, up to five photos per inspection, local saving, report preview and backup. That is enough for one real job from first finding to preview.
On the points above, the verified behaviour is this. Inspection records and photographs are saved locally, so capture continues without a connection, while purchases and periodic Pro verification need internet. Photos can be captured or imported, captioned, linked to a checklist item or finding and reordered. Measurements carry explicit units, and client, project and asset details are entered once.
Pro adds the custom checklist builder, photo annotation, up to 200 photos per inspection, company name and logo branding, and branded PDF or editable Word reports, at $9.99 per month or $79.99 per year.
Two honest limits. It does not connect to other software, and it has no team administration or seats, so it is built for one inspector rather than a managed workforce. It also does not write the report: findings and recommendations are the inspector's own words.
Your next step
Pick the job, book the time, and run the script on one app before you look at another. If you are testing for a portfolio of units rather than one inspector, the property manager guide covers what changes when the same inspection repeats.
