Efri Dereje

A HubSpot instance several years old, built by several people, none of whom still owned it.

This is the normal state of a HubSpot environment that has run for a few years without a dedicated owner. Workflows built for campaigns that ended long ago, still enrolling contacts. Lists duplicated because nobody could find the original. Properties created for one report and never used again. Nothing throws an error. Everything still runs. What breaks is anyone’s ability to answer a simple question about what’s working.

Audited the environment as it actually was before touching anything: every workflow, list, form, property and automation. What was running, what was running twice, what was running for no reason, and what people had quietly built workarounds for.

Then consolidation. Redundant workflows merged or switched off. Inherited automations that nobody owned got a decision made about them rather than being left to run. The environment got smaller before it got better.

Segmentation was rebuilt on how contacts actually behaved rather than on how the data had originally arrived, which is what makes everything downstream work: the sequences, the scoring, the reporting.

The result was a simpler, more maintainable HubSpot environment, with redundant automation consolidated, clearer segmentation and more reliable underlying data.

The team had a system it could better understand and reporting it could trust, rather than a growing collection of workflows and properties whose purpose was no longer clear.

Same company. Thousands of contacts built through the newsletter and social channels over several years. People who had already raised their hand.

Almost none of them were converting. Emails went unopened, and the few that were opened produced no action. Meanwhile paid spend was being increased to make up the shortfall, which meant the business was buying traffic to compensate for an audience it had already paid to acquire.

I redesigned the HubSpot lifecycle automation around buyer-journey stages and actual contact behaviour.

This involved rebuilding segmentation, enrollment and re-enrollment criteria, workflow branching, timing conditions, lead scoring and property updates. I also configured internal notifications to help sales identify contacts showing stronger buying intent.

The objective was to move away from sending the same communications to every contact and toward a system that responded to where people were in their journey.

I rebuilt the programme around buyer-journey stage. Segmentation first, then sequences mapped per stage rather than one linear path for the whole database. Subject lines rewritten. Message length cut substantially. Send timing changed. Personalisation built against what each segment had actually shown interest in rather than against a merge field.

Underneath that sat the workflow layer: enrollment and re-enrollment rules, branching by lifecycle stage and behaviour, delays and timing conditions, list membership and property updates, lead scoring, and internal notifications so sales knew when something was worth a call.

Nurture open rates lifted 40%. Outreach to the warm segments reached 81% open with a 9.26% reply rate. Those replies converted into signups.

A $67M software company operating in eight markets. Marketing budget and headcount had already been cut before I arrived in the role, because nobody could show what the function produced.

The business could see visitor counts. That was it. No connection between any marketing activity and any signup. Every budget conversation came down to opinion, and marketing lost those conversations because the other side of the table had numbers.

I owned the design and hands-on implementation of the measurement infrastructure connecting marketing activity to sign-ups.

This included developing a consistent UTM taxonomy across eight markets, configuring Google Tag Manager and GA4 conversion events, and implementing campaign and source attribution in HubSpot.

The objective wasn’t simply to produce more dashboards. It was to establish reliable data that the business could use to evaluate acquisition performance and make budget decisions.

A consistent UTM taxonomy across every channel and campaign, standardised so performance could be read per market rather than averaged across all eight. Google Tag Manager implemented hands-on: containers, tags, triggers, variables. GA4 event and conversion signals mapped to the actions that indicated genuine intent rather than browsing, which is the part most implementations skip. Then campaign and source attribution configured in HubSpot so a contact could be traced from first touch through to signup.

The last piece was a dashboard built for the people who had to make decisions from it, not for an analyst to interpret, with a reporting cadence that put it in front of them regularly enough to matter.

87% of marketing activity became traceable through to signups, from a baseline of nothing. 25 channels were measured for the first time.

The uncomfortable finding: roughly 80% of spend was going to channels contributing almost nothing. That budget moved. Legacy channels were cut, and one channel nobody had been looking at turned out to be producing the most engaged, purchase-ready traffic of anything in the mix.