Halo — compounding a B2B pipeline

Halo is a B2B data platform with strong product traction and a marketing function flying blind. They were spending well, generating leads, and yet nobody in the building could state — and defend — a single number for qualified pipeline. We rebuilt the measurement layer first, then let it drive every spend decision. Over nine months qualified pipeline grew 234%.

Overview

When Halo came to us, the problem wasn't volume. Forms were being filled, demos were being booked, and the sales team was busy. The problem was trust. Three analytics tools disagreed with each other, the CRM and the ad platforms told different stories, and the quarterly board deck was assembled by hand from spreadsheets that no two people built the same way. Marketing couldn't say which channels created revenue, so budget moved on instinct and the loudest opinion in the room.

Our engagement had one organising principle: make the pipeline number trustworthy, then make it move. Everything we built — attribution, scoring, reallocation — existed to serve a figure the VP of Growth could put in front of the board and stand behind without caveats.

The challenge

Halo's analytics were scattered across tools that never reconciled, and there was no trustworthy pipeline number anyone in the company actually believed. Demo requests were counted three different ways depending on who pulled the report. Paid channels claimed credit for the same opportunities, so the totals never added up. Sales discounted marketing's figures on principle, and marketing couldn't prove them wrong. Without a shared source of truth, every reallocation argument came down to seniority rather than evidence — and the spend stayed frozen in last year's mix.

Our approach

We didn't start with campaigns. We started with the ledger. The first six weeks were spent reconciling definitions and wiring a single pipeline of record, so that the rest of the work had solid ground to stand on. From there, prediction and reallocation could compound on top of clean data instead of fighting it.

Unify the truth

We consolidated every source — web analytics, the CRM, and each ad platform — into one attribution model with agreed definitions for a lead, an opportunity, and qualified pipeline. One number, one place, refreshed daily.

Score and forecast

On top of clean data we built a predictive lead-scoring model that ranked inbound by likelihood to become real pipeline, so sales worked the right accounts first and marketing could forecast pipeline weeks ahead of close.

• Unified multi-touch attribution across web, CRM and paid channels
• Predictive lead scoring trained on historical opportunity outcomes
• A weekly reallocation ritual moving spend toward proven sources
• A live dashboard tying every euro to qualified pipeline
• A structured experiment backlog, prioritised by expected lift

The results

Once the number was trusted, decisions got faster and braver. The weekly reallocation cadence meant underperforming spend was caught in days, not quarters, and freed budget flowed to whatever the model said was working. Experiment throughput more than tripled, because the team no longer argued about what happened — they read it off the dashboard and moved on to the next test.

Spira turned scattered data into a system the whole board trusts. Reporting went from guesswork to clarity — and the pipeline number finally compounds.

— Nina Costa, VP Growth