Laravel-based solution for an internal inventory tool
Discovery phase + technical blueprint for a B2B product directory. Laravel + Twill CMS, Algolia faceted search, Snowplow analytics, 40+ page architecture document.
I helped plan and build a B2B product directory utilising the Twill framework, Algolia search and Snowplow analytics.
The client's existing platform was end-of-life, with directory data management locked inside a third-party black box. The brief was a full rebuild - bring the data in-house, modernise the editorial surface, preserve every legacy search rule and SEO equity through the cutover.
I delivered the discovery phase, culminating in a 40+ page architecture document defining the technical requirements, system architecture, and a working proof-of-concept against the new stack. Scope expanded during the work as phase-one design surfaced cross-phase dependencies.
Four decisions the blueprint turned on:
Before any production code, I ran several discovery sessions with the client to understand exactly what the existing service did well, where it broke, and what the editorial team's actual day-to-day pain points were. The technical decisions - Laravel + Twill, Scout/Algolia, Snowplow - fell out of those conversations naturally. The architecture document was much sharper because it was aligned to the real requirements before a line of production code was written, and the PoC let the client see the new stack handling their actual data before committing to the full build.
The discovery phase ran smoothly. The client ended up with a working proof of concept that demonstrated every piece of the proposed system, plus the architecture document as a reference - a complete blueprint to take forwards on their own timeline.
Twill is genuinely underrated as a CMS base - for any client where Laravel makes sense but the editorial interface needs to feel modern, it deserves more attention than it gets.
All case studiesThis article was drafted with the help of AI to populate the page. I'm in the process of rewriting it - a principle I adhere to across all projects. AI produces boilerplate, not production-quality output.
Need a bespoke product catalogue with proper search + analytics? I love this kind of work.
Misbehaving stack? Codebase that won't play fair?