[ Case study ]
The food hall's site listed every vendor's hours, menus, and events, but updates went through the property group's overworked marketing manager; vendors emailed changes that went live late or wrong, and event weekends exposed stale hours that guests complained about publicly.
CLIENT a food hall operated by a property group — FOCUS Scoped editing per vendor
Representative examplesEvery case study in this library is an illustrative composite of the kind of engagement we deliver — written to show our method and standards, not to name clients.
Two dozen independent food vendors — counters, stalls, and a bar — share a restored warehouse under one roof and one brand, operated by a property group that also runs the website. The site is the practical tool guests rely on: who's open today, what's on the menu, what event is happening this weekend. Changes come from vendors by email to the property group's marketing manager, who re-keys them into the site alongside her duties for the group's other properties. Event weekends stretch that arrangement past its limit.
The food hall's site listed every vendor's hours, menus, and events, but updates went through the property group's overworked marketing manager; vendors emailed changes that went live late or wrong, and event weekends exposed stale hours that guests complained about publicly.
We proposed getting the marketing manager out of the re-keying business: each vendor receives an account scoped to their own listing — hours, menu, photo, description — with the rest of the site invisible to them. A light review queue would apply to the first month of edits while trust calibrated, after which proven vendors publish directly; the design and legal pages stay locked to the property group. A weekly maintenance rhythm — updates, backups, a monthly report — keeps the platform boring, which is the entire point.
Just as important is what we ruled out, and why:
Each vendor gets an account limited to their own listing — hours, menu, photo, and description — with the rest of the site invisible to them.
Vendor edits enter a light review queue for the first month, after which trusted vendors publish directly — control where it's needed, speed where it's earned.
Updates, backups, and a monthly report keep the platform boring, which is the point.
Delivered by the systems pod — engineer over 4 weeks, with working increments reviewed with the client every week.
Obstacle
The blanket review queue held every vendor's edits for a month, and the most confident vendors soured on the workflow within two weeks — the queue was protecting the wrong end of the spectrum.
Handled: We switched to trust earned from edit history: confident vendors published directly after week one, the review queue stayed for new accounts, and the manager reviewed exceptions monthly.
Obstacle
The least technical vendor's first self-serve edit flattened their listing's layout the day before a holiday weekend — the exact failure the property group had warned about.
Handled: We restored the listing from revision history in minutes, added layout-locking to the scoped editor, and scheduled a quiet-hours onboarding session for the vendors who wanted one.
The headline: vendors maintaining their own listings, with the marketing manager out of the day-to-day — Manager bottleneck → scoped vendor edits, read from Listing update timestamps. A second check: median hours-update turnaround at 9 days → same day.
The marketing manager's week changed shape: vendor updates went from a daily interruption to a monthly report to skim. Vendors update listings between services and see the result live, which changed how they talk about the hall in public. Guest complaints about stale hours all but disappeared, and event weekends lost their particular dread. The property group got something it hadn't had before: a website that runs itself well enough to be boring.
The result was read from Listing update timestamps against the pre-engagement baseline over the stated window, with a guardrail check on median hours-update turnaround. Where platform-reported numbers and business outcomes differ, this record says which layer it is quoting.
What we would do differently
We would launch the review queue off by default for the three most confident vendors — the month-long blanket queue taught us trust levels we could have read from the first week's edit history.
[ Related service ]
[ Related builds ]
Shared host per-site preview deploysAll five brands live on Vercel with preview-deploy reviews and tested rollbacks
Bot buyouts protected on-salesSubsequent on-sale days completed with bots challenged at the edge and origin load flat
[ Next step ]
Next case study