Ray Breslin
5 June 2026 · 5 min read
In October 2026 a new marketplace for Australian businesses for sale went live. BizExchange is built for businesses listed by private owners and by brokers, with the asking price and the key figures shown up front, and it launched with more than 5,000 listings. Core Creations designed and built it.
This case study covers what the site needed to do, the technology behind it, how listings get onto it, and the design, search and performance decisions that shaped it. If you are planning a marketplace, a directory or any site where the data is the product, most of these decisions will apply to you too.
What the site needed to do
Buying a business usually starts with a lot of clicking: open a listing to find the asking price, open another to see whether the profit is published, then try to compare the two in your head. The core requirement for BizExchange was to make that comparison possible at a glance, across thousands of listings, without the buyer having to sign up or pay.
That set the shape of everything else. The listing data had to be structured enough to compare, the pages had to be fast on a phone, and the site had to be easy for search engines to understand, so that a buyer searching for a type of business in a particular place lands on a page that answers it.
The technology
We chose a stack that is fast to work in, keeps data in Australia and scales without a rebuild:
- Next.js 16 and React 19, using the App Router and TypeScript throughout.
- Vercel hosting, with the server code running in Sydney, close to the people using the site.
- Supabase for the Postgres database and for sign-in, by email link or Google, also in the Sydney region.
- MapLibre maps on OpenStreetMap data, using OpenFreeMap tiles, so buyers can search by area on a map.
- Stripe Checkout for payments.
- Resend for email.
- Cloudflare Turnstile spam protection on the site's forms.
- Vercel Web Analytics, which is cookieless.
Every change is tested automatically before it goes live: browser tests, accessibility checks, and page-speed budgets for both phones and desktop, so a change that would slow a page down or break a form is caught before anyone sees it.
How listings get onto the site
A marketplace is only as useful as its listings, so getting businesses onto the site had to be simple for every kind of seller.
- Private owners list directly, and every new owner listing is checked within one business day before it goes live.
- Brokers and agencies add listings by hand, by spreadsheet (CSV) upload, or by a feed from their listing software. The brokerage is named on each listing.
- Some broker listings are short factual summaries of businesses the broker advertises publicly. Each one names the broker and says its figures are as published by the broker, and they are refreshed regularly. Enquiries go straight to the broker, who can claim the listing for free or have it removed.
Design decisions
Figures up front
Every listing card and listing page shows the asking price alongside turnover, profit and the asking price to profit multiple. Where a seller has not published a figure, the site says "Undisclosed" rather than hiding the field, so buyers can see at a glance what they would need to ask about.
Market context on every listing
Each listing has a "Market context" section that compares its multiple with the median for its industry. A buyer looking at a cafe can see straight away whether its asking price is high or low against other cafes on the site, relative to profit.
Live data that refuses small samples
The same listing data feeds a public insights section with median asking prices and multiples by state and by industry, updated daily. A figure is only published when it is based on at least 20 listings, and the method is published alongside it. It would have been easy to show a number for every industry from day one. We chose not to, because a median from five listings looks just as authoritative on screen as one from five hundred.
The "What's my business worth?" calculator is built on those same multiples, so owners get a range drawn from real asking prices in their industry rather than a generic rule of thumb.
Direct enquiries, simple pricing
Enquiries go straight to the seller or broker. Buyers never pay and BizExchange takes no commission. Owners pay a flat fee to list and brokers list free, with paid options for extra visibility.
Dashboards for sellers
Sellers have a dashboard that shows views, saves and enquiries for each listing. Views are counted without cookies, in keeping with the rest of the site.
Built to be found
Search structure was planned from the start rather than added later. The site has pages for 36 industries, for each state and territory, for each industry within each state, and for price bands, so a search like "cafes for sale in Queensland" has a page that answers it directly.
There are 24 guides covering buying, selling, valuation, finance, legal steps and specific industries. Pages carry structured data (Organization, ItemList, Article and Offer), the sitemaps carry accurate last-updated dates so search engines know what has changed, and the site is open to both search engines and AI search crawlers, with an llms.txt file describing it.
Private and fast
The site uses only the cookies it needs to work. There are no advertising or analytics cookies.
To keep pages fast on phones, images are resized and compressed for each device, and only the first listing photo loads with priority. Listings without their own photos use stock images from Pexels, labelled "Stock photo" on the card and credited to the photographer, so nobody mistakes a stock image for the business itself.
What we would tell anyone planning a marketplace
- Decide which data is the product. For BizExchange it was price, turnover and profit. Everything else, from card design to the insights pages, follows from making those comparable.
- Design for missing data. Real listings are incomplete. Showing "Undisclosed" honestly is more useful than a layout that only works when every field is filled.
- Let the data improve with scale. Thresholds such as the 20-listing rule mean the insights become more complete as listings grow, without ever publishing a misleading figure.
- Automate the quality checks. Speed budgets and accessibility tests that run on every change keep a large site fast and usable long after launch.
- Plan search structure with the data model. Industry, state and price pages are easy when the listing model was built for them, and painful to add afterwards.
If you are planning a marketplace, directory or data-driven web app, our custom web app development team can help you work through the same decisions, or browse more of our case studies.