Barlows Caravan Park asked a simple question: should there be an app? This case study covers how that question was answered, what the answer turned out to depend on, and what got built. At the time of writing this is a scoping document and a working prototype rather than a launched app.
At a glance
| Client | Barlows Caravan Park |
| Type | Progressive web app, scoping and prototype |
| Households | Around 500 (roughly 300 static pitches, a couple of hundred seasonal tourers) |
| Identified annual saving | £2,250 to £4,000 in postage and handling |
| Status | Scoping document delivered, working prototype built |
| Work | Requirements, feature scoping, prototype build |
The starting position
The owner asked five things. What would an app do, who would use it, would it need downloading from the app stores, could she update it herself, and could it show people their billing.
Those are the right questions, and the honest answer to the first one depended on a number nobody had stated yet.
What changed the answer
The park is roughly ten times bigger than the website suggests. Around 300 static pitches and a couple of hundred seasonal tourers is about 500 households.
My first draft assumed 50 to 200 and said the case was reasonable. At 500 it stops being a judgement call. Nothing else in the scoping document mattered as much as that number, and it only surfaced because the owner came back with detail after the first draft. That is worth saying plainly: the first version of the answer was wrong because it was built on an assumption rather than a figure.
The diagnosis
Two of the three audiences are recurring. Static owners are there for years and seasonal touring owners come back season after season. Occasional tourers stay a few nights and will never install anything, and shouldn’t be asked to. An app earns its keep on the first two and needs to not obstruct the third.
Guests were already doing the app’s main job by hand. Storm and heavy rain warnings were going out on Facebook, and guests were notifying each other about awnings and damage. That’s not a hypothesis about what people might use. It’s an existing behaviour running on the worst possible tool for it. A notification reaches every touring pitch. A Facebook post reaches whoever happens to scroll past.
The information flow was the actual problem. Prices were hung on the shop wall, so you had to physically be at the shop to know what anything costs. Nobody knew what gas was in stock without walking down to ask. This wasn’t a park that needed an app to feel modern.
The postage number was the clearest hard saving. Two site fee invoices a year, plus electricity and annual rates, is three or four documents per household per year. Across roughly 500 households that’s somewhere between 1,500 and 2,000 posted items annually, at a pound and a half to two pounds each once you count paper, ink, envelopes, postage and the time spent stuffing them.
What the detail changed
Two findings reshaped features rather than confirming them.
Groundworks on the park are constant rather than occasional. A dated post that scrolls away is the wrong format for “the lower road is single width for a fortnight”, so the prototype carries a standing Works on the Park block that sits above the notices and stays until the work finishes.
The closed season turned out to be an argument for the app rather than against it. Open 1 March to about 6 January means roughly eight weeks when 300 owners are away and can’t check their own caravans, in the windiest part of the year. A message saying the park has been walked after last night’s gale, and here’s what was found, is worth more to an absent owner than anything the app does in August.
The feature test
Every feature had to remove a phone call, save the owner a job, or give someone a reason to open the app more than once a year. Anything failing all three stayed out.
That test did more work than any feature list would have. It’s also the part clients tend to find useful afterwards, because it survives the project.
The risk that closed
The park has no wifi beyond the shop, and owners run their own Starlink. That would have been a serious problem, because a notification can’t reach a phone with no connection.
It turned out mobile coverage is good across the park on all the networks, so notifications will land. The app is still built offline first so cached content stays readable, but the risk flagged in the first draft is closed.
Billing
Accounting software made the billing question answerable. Payment arrangements vary, with some owners paying monthly and some in two instalments, so the account section shows a schedule and what’s next rather than just a list of invoices. “What do I owe and when” is a different question from “send me my invoice”, and only one of them is answered by a PDF.
What I learned
Scope from the behaviour that already exists, not from the feature list. The single most useful thing in everything the owner told me was that guests already warn each other about storm damage on Facebook. Every strong feature in the prototype traces back to something people were already doing awkwardly.
The other learning is about being wrong early. Publishing a first draft with its assumptions visible is what got the real household numbers out of the client, and those numbers changed the recommendation. A more confident first draft would have been a worse one.
Limitations
This is a scoping document and a prototype. It hasn’t launched, so there are no adoption or engagement figures to report, and the postage saving is a calculated figure rather than a measured one. The prototype is live for review, and the full build is waiting on the owner’s sign-off.
Services used
Progressive web apps and app and tool development. For the park’s wider search work, see the Barlows Caravan Park case study and the local businesses and communities page.
FAQ
Does a caravan park need an app?
It depends on how many recurring households there are and how information currently reaches them. At around 500 households, with constant works, real weather exposure and an invoice run going out by post, the case is strong. At 50 households it would be much harder to justify the spend.
Would people actually install it?
The ones who come back will, if it does something they need. The test I use is whether a feature removes a phone call, saves the owner a job, or gives someone a reason to open it more than once a year. Occasional visitors won’t install anything, and shouldn’t be asked to.
Can the owner update it herself?
Yes, and that was a requirement from the first conversation rather than an afterthought. Notices, works updates, prices and gas stock all need to be editable without a developer, because otherwise the app stops being current within about a month and people quietly stop opening it.
How much does posting invoices really cost?
For this park, between £2,250 and £4,000 a year. That’s 1,500 to 2,000 items at a pound and a half to two pounds each, counting paper, ink, envelopes, postage and the time spent stuffing them. Most businesses only count the stamp.
Got a community of repeat customers?
Get in touch and tell me how you currently reach them. That’s usually enough to say whether an app is worth the money.