Redesign Health

Growth Today rebuilt how Redesign Health identifies and reaches founder archetypes, using AI-driven multi-source enrichment and a scoring layer, then deployed the same go-to-market play to the portfolio companies.

Impact TL;DR

• A seniority and function matrix producing 250+ qualifying title variations and 70 suppression keywords
• Multi-source AI enrichment that cross-references providers rather than trusting a single one
• A scoring and disqualification layer applied before a contact enters a sequence
• A custom reporting dashboard built because native campaign reporting was too coarse
• Hypothesis-driven messaging tests on what founders actually value in a backer
• The method extended to portfolio companies as a second ICP on the same system

‍

Adjacent job titles that describe completely different people

Redesign Health's targeting needed a level of precision that off-the-shelf filters do not offer. In their security-adjacent segments, a cloud security engineer and a network security engineer are both in scope while a network engineer is explicitly out. Dozens of distinctions like that existed across the target profile, and a single filter list could not express them.

The sourcing side had a second problem. Redesign Health wanted to develop its own view of the ideal founder and partner archetype, which is a research question before it is a prospecting question. Their existing view was qualitative, and there was no quantitative method to test it against.

Reporting was also too coarse to manage. The native campaign dashboard could not show performance at the granularity needed to spot deliverability problems or messaging angles worth testing.

Key challenges were:

• Adjacent titles requiring inclusion and exclusion at keyword level, not category level
• Founder archetypes defined qualitatively, with no quantitative test
• Single-source enrichment insufficient for a technical, nuanced ICP
• Native reporting too coarse to catch problems at segment level
• Two different ICPs in one organization, with different motions

‍

Growth Today built the targeting, enrichment and measurement layers, then treated messaging as a set of testable hypotheses

Instead of filtering on job title categories, we built a system where:

• Inclusion and exclusion are expressed at keyword level across a full title matrix
• Enrichment cross-references multiple sources and uses AI to reconcile them
• Scoring and disqualification run before a contact reaches a sequence
• Reporting is granular enough to catch a problem in one segment
• Messaging tests answer a stated hypothesis about what the audience values
• The same method serves both ICPs, sourcing and portfolio

None of this was going to work without a shared definition of the target. So the engagement opened with three structured workshops: a GTM workshop on positioning and differentiation, an ICP workshop on segmentation criteria, and a copywriting workshop on the prospect's own motivations.

‍

Generating the target titles from a seniority and function matrix

What we did: Rather than list target titles, we built a matrix of seniority levels against functions and generated every plausible combination and variation from it. That produced 250+ qualifying job titles. We then built 70 suppression keywords for the adjacent titles that look like matches and are not.

The outcome: The targeting expresses distinctions a filter cannot, like including cloud and network security engineers while excluding network engineers. Precision came from the exclusion list as much as the inclusion list.

If your ICP has technical adjacency problems, build the suppression list with the same care as the target list. The near-misses are what consume a campaign.

‍

Enriching from several sources and reconciling with AI

What we did: We built enrichment that cross-references multiple providers for each record rather than accepting one source's answer, then uses AI to reconcile conflicts and fill gaps. Provider integrations and AI retrieval were each used where they were actually stronger, since integrations cost credits and AI retrieval costs accuracy in different places.

The outcome: Higher-quality CRM records for a technical ICP where single-source data is frequently wrong or stale. A scoring and disqualification layer then ran on top, filtering records before they entered a sequence.

‍

Building the reporting the native tools did not offer

What we did: We built a custom dashboard for campaign statistics because the native reporting was too limited. It refreshes daily and also updates via a webhook whenever a new response arrives, so the numbers reflect the current state rather than the last scheduled pull.

The outcome: The team can see performance segment by segment, which is what lets a deliverability or messaging problem get caught in one segment instead of after it has run across the programme.

‍

Turning messaging into stated hypotheses

What we did: Rather than write copy and measure it, we built messaging tests around explicit hypotheses about what a healthcare founder values in a venture partner. Candidate hypotheses included regulatory support through FDA processes, the strength of the backer's healthcare network, and the expectation of follow-on funding through long clinical trial cycles. Each became a testable variant.

The outcome: Redesign Health can now say which pitch to lead with when approaching a founder, backed by test results rather than preference. We also set expectations on what the tests require: these comparisons only produce reliable answers at volume, since a low-volume variant yielding one or two responses does not support a conclusion.

‍

Extending the method to the portfolio

What we did: The second ICP in the engagement was buyers for the portfolio companies rather than founders for the studio. The same title matrix approach, enrichment layer, scoring model and reporting were applied to portfolio company go-to-market rather than rebuilt per company.

The outcome: Portfolio companies inherit a targeting and enrichment method that has already been validated at the studio level, instead of each one solving the same problem from scratch with a different vendor.

‍

Why Redesign Health Chose Growth Today

Comfort with a sophisticated buyer and an unusual motion

Redesign Health's team is technically sophisticated, and the ask was not standard outbound. Sourcing founders is close to the reverse of the fundraising motion most outbound teams have run. They partnered with Growth Today because they wanted to reverse engineer that logic with a transparent, reliable partner who could also advise on strategy, not only execute.

• Precision targeting. A title matrix and suppression list built for technical adjacency
• Multi-source enrichment. Cross-referenced providers with AI reconciliation, not a single source
• Knowledge transfer. The engagement was scoped so the internal team builds the muscle

‍

Results & Impact

RevOps Impact

• The team targets on 250+ qualifying titles and 70 suppression keywords, generated from a seniority and function matrix rather than written by hand
• Every record is checked against several providers, with AI reconciling the conflicts
• Bad-fit records never reach a sequence, because scoring and disqualification run first
• The team sees segment-level numbers the same day, refreshed daily and on every new response

Sales Impact

• The team can test what founders actually value in a backer, one stated hypothesis at a time
• One system runs both motions, founder sourcing for the studio and buyer outreach for the portfolio
• Reps stop working near-miss titles, filtered out before the sequence rather than after the reply

Leadership Impact

• Leaders can say which pitch to lead with when recruiting founders, tested rather than assumed
• Leaders can hand a validated sourcing method to a new portfolio company instead of paying to rebuild it
• Leaders can define the founder archetype quantitatively, and revise it as the tests come back

By generating the target matrix instead of listing it, Redesign Health built founder sourcing that its portfolio companies can inherit.

‍