If a person must finish a job where mobile coverage is unreliable, the essential part of your field app should work without a connection. That means more than opening yesterday’s schedule. The person must be able to record work, keep it safely on the device, see what has not synced and finish the handoff when service returns. Test those steps before you buy or commission the software.
You may decide that a connected app is perfectly adequate for your business. A team working mostly in an office or on reliable Wi-Fi may have little reason to pay for the extra design, testing and support that offline work requires. The decision belongs to the job, the sites and the cost of a failed visit.
Where does the job actually lose its connection?
Ask the people doing the work to mark the points in a normal week when the app becomes unreliable. A workshop, basement, plant room, rural site or vehicle can all present different conditions. Test the devices and network providers your team actually uses. A coverage map is useful for planning; it cannot tell you whether a particular room will load a form at the moment it is needed.
Ofcom’s Connected Nations 2025 report illustrates the distinction. Its modelled 4G coverage on B roads was lower inside vehicles than outside them across the operators measured. Those figures are about roads, not your premises or a forecast of failed jobs. They are a reason to test the real working environment instead of treating a general coverage claim as acceptance evidence.
Write down the job that must continue: receiving an assignment, finding the site details, taking a reading, attaching a photo, obtaining a signature, completing a certificate or recording materials. Not every function needs offline support. Sending an invoice to a customer, checking live stock at another branch or taking payment may reasonably wait for a connection. Be explicit about what must happen now and what can queue.
A vendor may advertise “offline mode”. That phrase does not establish which records are available, how long they remain on the device or what happens when the app restarts. VioTrade’s October 2026 release notes list jobs, notes, photos and other tasks it says work without signal. That is a useful example of a vendor making the scope visible, though I have not independently tested its app. Ask any supplier for the same level of detail, then try it with your own workflow.
What should the app let a person do offline?
Start with the smallest complete job. For a service visit, the person may need the address, contact, access instructions and previous equipment details before leaving the office. At the site they should be able to add readings, materials, notes and photos, then mark the job ready for review. If a required item is missing, the app should say so while the person can still correct it.
Separate the stages in your test:
| Stage | What to prove |
|---|---|
| Before travel | The assigned job and necessary reference details are on the device, not just visible while online. |
| At the site | New entries and attachments save with no signal and remain after the app is closed and reopened. |
| Waiting to sync | The user can see which changes are local, queued, failed or confirmed by the server. |
| Back online | Each change reaches the correct job once, with attachments and timestamps intact. |
| Office review | The office can inspect the completed record and handle any exception before invoicing or filing. |
The app should make uncertainty visible. “Saved” may mean saved locally, accepted by the server or passed to the accounting system. Those are different states. If the interface uses one green tick for all three, ask how a user can find a change that is still waiting or has failed.
A field worker should also be able to tell when reference information is old. A job downloaded yesterday could have been cancelled or rescheduled while the device was offline. Show the last update time and give the team a clear rule for checking changes before starting sensitive work. No software can magically supply a new instruction to a disconnected phone.
How do you test the claim before signing?
Ask the supplier for a trial that permits real offline testing. Use safe, representative records rather than customer data you do not need to copy. Include the awkward cases that a product demonstration tends to skip:
- Prepare a job online. Include a site note, an attachment and a required field. Confirm the assigned device has received them.
- Disconnect the device completely. Turn off Wi-Fi and mobile data. Open the job, change a reading, add a photo and finish a note. Do not accept a demo that merely shows a previously opened page.
- Close and restart the app. Confirm every change is still there. Restart the device as well if losing a day’s work would be material.
- Make a conflicting edit. While the first device is offline, let an authorised office user change the same job online. Reconnect and inspect which values win, which are flagged and whether either edit disappears.
- Interrupt the upload. Reconnect briefly, disconnect during a photo upload, then reconnect again. Check whether the attachment arrives once and the job can be reconciled.
- Follow the downstream handoff. If a completed job creates an invoice, certificate or stock movement, check that it happens once and that a failure has a visible owner.
Record expected and actual outcomes. A pass means the business record is correct and a user can finish the task, not simply that a sync indicator stopped spinning. If the supplier cannot provide a suitable trial, ask for a witnessed demonstration using these cases and treat anything you cannot inspect as an open question.
You should also ask how long queued work remains safe, how much local storage it needs and what happens if the device is lost before syncing. Those answers may depend on the phone’s operating system, security settings and the product’s design. The supplier should explain its limits and recovery path in terms your team can use.
What does “offline first” add to a software build?
In a connected app, the server can usually validate and save an action immediately. An offline app must keep enough data and rules on the device to let work proceed, then reconcile the result later. MDN’s guide to offline and background web operations explains some of the browser mechanisms. Caching a screen through a service worker is only one part of the job. It does not decide who owns a changed record or repair a failed invoice handoff.
For a bespoke field product, I would specify the record and sync rules before choosing a mobile framework. Which job details are copied to the phone? Who can see them? How are attachments stored and removed? Can two people edit the same field? How do we identify a repeated upload so it does not create a second invoice? What can a user do when a change needs attention? These are product and operational decisions, not settings to leave until the last sprint.
Offline support also creates maintenance work. New fields, validation rules and server changes must work with older app versions that have queued records. Someone needs to monitor failed syncs, support devices and test updates under poor connections. Include that work in the running cost of a bespoke option. Our mobile app development service can scope those behaviours around the actual field job rather than promising every screen will work offline.
Should you buy, configure or build?
Buy an existing product when its supported offline functions cover the important job and your team can operate its exception process. Configuration may be enough if the product already has the records and sync behaviour you need, but its forms or permissions need adjusting. A focused bespoke component can make sense when an important rule or handoff is distinctive and the available products cannot pass your test. That may be a field capture tool connected to an existing accounting or job system, rather than a replacement for everything.
Compare the full options, including subscription and device costs, implementation, integration, training, support and future changes. Add the business effect of a failed visit or duplicated record if you can estimate it from your own evidence. Mark unknown values as unknown. An inexpensive licence can be good value when it works; a custom build can be good value when it solves a material problem and has an owner for maintenance. Neither conclusion follows from the label “SaaS” or “bespoke”. Our build versus buy guide sets out the wider decision.
Start with one workflow and a clear pass condition. If the team cannot finish that job safely without signal, decide whether a supplier can close the gap, whether the process can wait for connectivity, or whether a bounded build is justified. Do not pay for offline support across the whole product just because one site occasionally has a dead spot.
What should go into the acceptance note?
Keep one page alongside the buying decision. Name the job and the people who use it. List the actions that must work without signal, the reference data that must be on the device, and the maximum acceptable delay before the office sees the result. State what happens to conflicting edits, failed uploads, lost devices and downstream records. Include the test devices, locations, dates and actual results. Name who will investigate failures after launch.
That note is useful whether you buy or build. It gives a vendor a specific claim to demonstrate and gives a developer a result to deliver. It also gives your team a reason to reject a polished demo that leaves a real job unfinished.
Recommended reading
If you are deciding whether a field workflow needs a different product or a small custom component, tell me which job fails and where it happens. I can help turn that into a testable scope before you commit to a build.