About 20 days from first call to live site, which is roughly three weeks. Here is the day-by-day plan: what happens, where projects usually go wrong, what needs the dealer's time, and where the days actually go. I'm publishing this because "custom build" makes dealers picture six months of silence followed by an invoice, and the honest answer is that a single-rooftop independent site is a matter of weeks — but only if specific things happen on specific days, mostly involving the dealer.
This is a planning timeline for a typical single-rooftop independent, not a record of one client's project and not an industry average. My own FAQ currently says timelines depend on scope and get mapped out on the first call, which is true and also a non-answer. This is the answer.
Why publishing this is worth doing
A day-by-day plan exposes your estimating. If I say the photo import takes three days, you can hold me to it. I'd rather a prospective client read the real shape of a project — including the parts that depend on them — than sign up expecting a process that doesn't exist.
It also settles the objection nobody says out loud, which is what happens if the one developer disappears. A public plan of how the work is structured, where the code lives and how it's documented is a more useful answer than a reassurance.
Week 0: before any code
Before quoting anything, I need answers to a handful of questions. A dealer who prepares these before a first call, with me or anyone else, will get a more accurate quote and a faster project:
- How many vehicles are in stock, and how quickly do they turn?
- Where does the inventory data live: a DMS, a feed provider, a spreadsheet?
- Who updates the site day to day, and from what device?
- Do you need pages in more than one language?
- What is the current site, what does it cost per month, and when does the contract end?
- Which fees apply to every sale, so the price can be displayed correctly?
- Which online steps, such as credit applications, F&I products or price finalisation, are in scope for version one?
One scoping note worth making generalisable. Cox Automotive's 2025 study found consistent gaps between what buyers want to do online and what they actually can: 48% want to apply for credit online but only 33% do; 40% want to select F&I products online but 16% do; 37% want to finalise price online but 19% do. Those gaps are where a dealer's build budget has the most room to earn its money — and also where scope inflates fastest. Deciding which of them you're tackling in version one, before design starts, is the single highest-leverage hour of the whole project.
Day-by-day
Each stage below covers what happens, what tends to get blocked, and what needs the dealer. The day counts are planning estimates for a single-rooftop independent, and they move with scope and with how quickly the dealer responds.
Days 1–2: Design direction. Not a full mockup, just enough to agree the shape: homepage, inventory listing, vehicle detail page and contact. The dealer sees real layouts, not a description.
Day 3: Design approval gate. The project does not proceed past here without written sign-off from one named decision-maker. Anything requested after this point goes on a version-two list unless the dealer accepts the date moving.
Days 4–6: The vehicle page first. The most important page gets built before the homepage, because that's where buyers land and where the money is.
Days 7–9: Inventory import and the data reality. Every build has this stretch: the real data is messier than the sample. Expect wrong trims, missing options, 6 MB phone photos and inconsistent mileage formats, and expect to write normalisation rules for all of them. The inventory plumbing article covers how that data gets from the lot to the page.
Days 10–12: The CMS. Built around one constraint: the owner updates stock from a phone, standing in the lot. The first version always has more in it than the owner needs, and the useful work is removing what they never touch.
Days 13–15: Forms, tracking, consent, security. Server-side validation, admin hardening, consent-gated analytics and backups. These are the things that get skipped when a project runs late, which is why they're scheduled rather than optional.
Days 16–18: QA. A keyboard-only pass, a screen reader pass, PageSpeed on the vehicle page rather than the homepage, cross-device checks, end-to-end form submission, and the sold-vehicle path.
Days 19–20: Launch and the 48 hours after. DNS, a redirect map from the old URLs, sitemap submission, Search Console, and watching the indexing report.
Where the days actually go. On projects like this, design and code tend to be the minority of the calendar. Data normalisation and waiting on input from the dealer usually take the majority, which is why the dependencies below matter more than anyone's typing speed.
Where builds usually go wrong
The photo import takes longer than planned. Phone photos arrive at full resolution, in inconsistent order, and sometimes with another vehicle's photos mixed in. Budget extra days for it, and build the resize-on-upload pipeline before the import rather than after.
A scope request arrives after design approval. A finance calculator, a trade-in form or a second language becomes "just one more page." The approval gate on day 3 exists for this: new requests go on the version-two list unless the dealer accepts a later launch.
A feed field doesn't mean what its name suggests. A "price" field that holds a pre-fee number, or a "status" field where "available" includes cars on hold. The defence is comparing a sample of imported vehicles against the actual cars and the deal paperwork before anything goes live.
What needed the dealer, and when
Here are the dependencies, in the order they're needed. The single best reason a build overruns is waiting on one of these:
- Day 1: answers to the discovery questions, logo and brand assets.
- Day 3: design approval from one named decision-maker.
- Day 7: DMS or feed credentials and a real inventory export.
- Day 10: the complete list of fees that apply to every sale, for price display.
- Days 10–12: the owner tries the CMS on their phone and reports what's awkward.
- Day 16: sign-off on final content: hours, staff, policies and copy.
- Day 19: domain registrar access for the DNS change.
What a platform migration looks like by comparison
Being straight about it: a template platform can have you live faster than this, because the design and infrastructure already exist and the work is configuration and data connection. If speed-to-live is your only criterion, that's the honest recommendation.
What you're buying with the extra weeks is a site shaped around your inventory and your process, with no monthly rent and code you own. Whether that trade is right for your dealership is the subject of the custom versus platform comparison, and the cost side is in what a dealership website actually costs.
What this means for your dealership
If you're getting quotes, ask any developer or vendor for a day-by-day plan with the client dependencies marked, and ask what happens on the day the real data turns out to be messy. Those two answers separate people who've done this from people who haven't.
And prepare your side before you start: feed or DMS credentials, domain access, your complete fee list, brand assets, and one named decision-maker who can approve a design. Projects don't slip because code is slow. They slip waiting for one of those five things.
Where I'd tell you not to hire me
A three-week timeline is realistic for a single-rooftop independent with a manageable inventory and no OEM obligations. If you're franchise, multi-rooftop, or need a library of DMS and CRM integrations, the honest timeline is much longer and a platform that already holds those integrations will beat me on both time and cost. I'd rather say that on the first call than discover it on day 30.
Sources
- Author's experience of how single-rooftop dealer builds run. Day counts are planning estimates and vary with scope and client response time.
- Cox Automotive, Car Buyer Journey Study (13 Jan 2026) — online-step gaps between what buyers want and what they complete.
Want the day-by-day plan for your own project before you commit to anything? Ask me for it.