INV—W/ANA · CASE · 2024 ANALYTICS · BUBBLE → DIRECTUS · FIG. 01
Invental/ Works/ Analytics Platform
Analytics Platform Low-code IT Incident Management NDA engagement

Analytics for IT incidents & ROI.

A two-phase low-code build for a leading IT service provider — migrating the analysis engine from Bubble to Directus while keeping the Bubble frontend, so the platform scales past 100,000-row files without blinking.

ConfidentialNDA
Product build · Low-code · Re-architecture
6 months · Two phases
Frontend (Bubble) · Analysis engine (Directus) · Integrations
— Client URL withheld
IT services · Enterprise analytics
— Workspace/Incidents/Overview
Live · last 24hv2.0.3
MTTR — mean time to resolve 28m ▼ 34% vs last period
ROI — estimated 312% ▲ projected · fy'24
Incidents — open 147 ▲ 12 new · p1
Rows ingested · last run 124,8k ▲ directus engine
— Incident volume vs resolved · trailing 90 days
WeekMonthQuarter
Resolved
Opened
P1 · critical
W01W04W08W12
— Active incidents24 open
▨▨▨▨▨ ▨▨▨ database lag— SEV-1 · region ap-east
02m
Auth service ▨▨▨▨▨▨▨ 5xx— SEV-2 · gateway-prod
14m
Ingest pipeline queue ▨▨▨▨— SEV-3 · directus-api
41m
▨▨▨▨▨▨ rate-limit trips— SEV-2 · edge-cache
1h
▨▨▨▨ ▨▨▨ certificate renew— SEV-3 · tls-mgr
2h
Payments ▨▨▨▨▨▨ spike— SEV-1 · billing
3h
— Ingest stream · directus engine124,832 rows · 00:00:04.2
[ FIG. 01 ] — Conceptual dashboard · representative only INV-W-ANA-2024 NDA · visuals are renders
[ M · 01 ]
100k+
Rows per file processed by the new Directus-powered analysis engine
[ M · 02 ]
6mo
End-to-end delivery across two phases — v1 on Bubble, v2 on Bubble + Directus
[ M · 03 ]
2×
Stacks working in concert — Bubble for UI, Directus for data
[ M · 04 ]
1
Unified analytics surface for incident management and ROI

§ 01Context

Our client — a leading IT service provider whose name is under NDA — runs incident management across a large, multi-region IT estate. They needed an advanced analytics platform that could both surface what's happening across that estate and translate it into an ROI story for the business. The original build sat entirely on Bubble; it worked, but it buckled under the data volume they actually operate at.

§ 02Challenge

Bubble is wonderful as an application layer, but it's not a data warehouse. The initial platform struggled with files over 100,000 rows — analyses that should run in seconds were timing out, and the ROI calculator couldn't be trusted on real-world exports. The brief was sharp:

Bubble for the surface. A real database underneath. The user shouldn't feel the seam.— Engagement kickoff

§ 03Solution

We delivered a two-phase build. Phase one stayed entirely on Bubble so the client had a live platform to iterate against. Phase two carved out the analysis engine and rebuilt it on Directus — an API-first data platform — then re-wired the Bubble frontend to call it. The seam is invisible; the speedup isn't.

What Directus changed

§ 04Versions

Delivered in two phases so the client had working software from month one, and a scalable platform by month six.

Phase 01 · proof of value
v1.0

Entirely on Bubble

Frontend and backend both on Bubble. Fast to ship, good for smaller datasets, perfect as a proof of concept for the client team to validate the analytics model.

Bubble · UIBubble · DBProof of concept
Phase 02 · scale build
v2.0

Bubble frontend · Directus engine

Data processing moved to Directus — an API-first data platform built for volume. Bubble continues to own the interface; Directus does the heavy lifting behind it.

Directus · engineBubble · UIREST · GraphQL100k+ rows

Same UI. New spine. 100k+ rows in seconds, not minutes.

— Phase 02 close-out

§ 05Result

The upgraded platform hit all four outcomes the engagement set out to deliver:


— End of case. Low-code hitting a ceiling? Let's talk about a migration path.

Architecture, end to end.

[ FIG. 02 ] · Phase 02 · Bubble frontend calls a Directus API-first engine backed by PostgreSQL
— Surface
Frontend Bubble · web app Dashboards · filters · ROI calculator UI · user management
Users Analysts · ops leads Run incident analyses · pull ROI reports
— Engine
Analysis engine Directus · API-first REST + GraphQL · row-level perms · background jobs
Integration Bubble ↔ Directus Signed API calls · schema-aware data contracts
Inputs CSV · API feeds · exports Incident logs > 100k rows · ticketing exports
— Data
Datastore PostgreSQL Indexed tables · fast aggregation · historical series
Processing Batch · on-demand Heavy transforms run server-side · frontend stays thin
API calls · Bubble ↔ Directus
Hot data path
Batch ingest · source feeds
The migration was invisible to our analysts and transformative for our ops team. We can finally run ROI on the real dataset — and the numbers hold up.
Director of IT Operations · client under NDA
— Engagement close-out · attributable on request
— More work View all case studies ↗
Promoo.io
2023Low-code
Promoo.io — promo codes platform
Besample
2024Low-code
Besample — research recruitment
Vitrina
2023Low-code
Vitrina — multivendor marketplace

Low-code hitting a ceiling? We do the re-architecture.

Bubble, Webflow, or another builder carrying more than it was designed to carry — we move the heavy parts to a real backend without rebuilding your UI.

Book an intro call