I Compared Building a Car Classifieds Site From Scratch vs Going With a Ready-Made Solution. Here's What Actually Happened.

Four months. Two developers. Still no working search filter. That's where I was when I almost gave up on the “we'll just build it ourselves” plan for a car classifieds marketplace.
Not a side project either — real budget, a real deadline, and a founder who wanted listings live before the next used-car buying season started. The decision came down to two paths: build it from scratch, or start from a ready-made base and customise from there. I ended up living through both, and the gap between what I expected and what actually happened is worth writing down.

The Build-From-Scratch Path
We started the usual way. Mapped out the core features — listings, seller verification, search filters, image uploads, in-app chat between buyers and sellers. Hired two developers. Set a timeline of ten weeks for a working version.
Four months in, we still didn't have a working search filter. Not because the developers were bad — they were solid — but because nobody had accounted for how many small decisions a car marketplace actually needs. Do you filter by VIN? By dealer versus private seller? How do you handle a listing where the odometer reading looks off? None of this shows up in a feature list. It shows up in week six, when someone finally tries to use the thing.
Seller verification alone took three weeks longer than planned. The team had never built an SMS OTP flow before, and the first gateway we tried had a delivery rate bad enough that test users assumed the app was broken. Small problem, big delay.
By the time we had something demoable, we were about 60 percent over the original budget estimate and nowhere near launch.

The Ready-Made Path
At that point we tested a different approach on a parallel track — evaluating existing car classified marketplace solutions instead of writing everything from zero. Spent about a week trying three of them with dummy data before picking one.
The one we picked already had VIN-based search and dealer inventory management built in — two things our custom build was still struggling with in month four. Real listings were live in twelve days. Not perfect, not fully branded, but working, and buyers could actually search by make, model and year without anyone writing a line of filter logic.
It wasn't plug-and-play in the way the sales page implied. Getting the branding right, adjusting the seller onboarding flow, and removing features we didn't need took another two weeks before it felt like our product rather than someone else's template. That part is easy to underestimate going in.

What the Numbers Actually Looked Like
Rough comparison, from what we tracked: the custom build was on track to cost close to four months of two developers' time before reaching a usable version. The ready-made route got to a working, listable marketplace in under a month, with the remaining budget going toward customisation and marketing instead of core plumbing.
That doesn't mean custom-built was the wrong call in general. If the business genuinely needs something nobody else has built — a matching algorithm specific to your market, a compliance flow unique to one country — writing it from scratch still makes sense. Most car classifieds businesses, though, are not solving a new problem. They're solving a well-understood one, slightly differently.

What I'd Tell Someone Starting This Decision Today
Good ready-made marketplace bases already handle the parts that quietly eat months on a custom build: VIN and dealer search, seller verification, listing management. That's not a small thing when the real constraint is time to first listing, not lines of code.
The honest lesson from this one: list the three or four features that would actually be hard to find off the shelf. Short list? A ready-made base plus customisation will almost always beat starting from zero. Long list? Budget for the four months, not the ten weeks. Either way, know which list you're on before you write a single line of code.