Smart Prospecting for B2B Sales Teams

In outbound sales, reps research every target company by hand across many sources to judge whether it's worth pursuing. Goava's first platform, “Goava classic”, pulled that company data into one place, sent news alerts on followed companies and recommended prospects based on existing customers.

But it had a churn problem. Once customers felt they knew their market, they stopped using it, and even with all the data in one place, users still had to do a lot of searching. The answer was Goava Discover: a new platform built around smart recommendations, prospecting, insights and advanced filters.

We ran interviews with 5 groups of 3-5 users each: Goava's own sales team plus existing and new customers across the Nordics, guided by our NPS survey data.

1. Every team has a pipeline, and CRMs skip the first step. 100% of users worked a pipeline of Discovering → Qualified → Converted/Denied → Customer. 76% used a CRM, but no CRM helped with the discovering stage, which left room for Discover.
2. Users came for speed. 83% used Goava to save time, and 94% needed the main facts about a prospect in one view: company details, key people, financials, location and a clear answer to “should I pick this company?”
3. Everyone needed CRM sync. 100% said integration mattered, because they didn't want their tools out of sync.
4. Users follow company news. 82% needed to follow news and events on the companies they track, both customers and prospects.
5. Needs vary by team. 47% wanted ownership data and 32% needed web and technology data, so one fixed layout wouldn't fit everyone.

Vainu and Crunchbase were the closest alternatives. Vainu covered the Nordics with solid search; Crunchbase had strong filters and financial data but was US-focused. Neither offered smart prospecting or recommendations.

1. Goava Score. A 1-100 score for how closely a company matches your existing customers. It updates as customers are added and as companies move through the pipeline, so every qualify or reject makes the next recommendation better.
2. Goava Fit. The most important reasons for and against a prospect, placed where you can't miss them. Generated by machine learning trained on each customer's own customer and pipeline data.
3. Three views for three ways of working. Dashboard for looking at one company in depth, Thumbnail for quickly working through many, and List for bulk actions and import/export.
4. Modular filters. Goava has a lot of filters but each user needs only a few, so users add just the ones they need and save them. A relevance sort uses machine learning and customer lists.
5. CRM first. Apps in the major CRM marketplaces, Zapier for the rest, and spreadsheet upload for customer lists, so recommendations start from real data on day one.

Discover would be iterated on constantly, so I built the Goava Design System alongside the high-fidelity designs: foundations first (color, spacing, typography, grid), then components, following atomic design. The main flows met WCAG 2.1. The same system now powers Goava's AI features, including the Prospecting Agent.

Discover is entirely data-driven, so Figma mockups alone couldn't test it. We built a limited MVP front end with real company data, plus small Figma prototypes for individual features, and tested with groups of 3-5 users.

• 86% found the recommendations useful for prospecting quickly
• 78% said the Goava Score helped them pick prospects
• 93% were confident the system would work for them
• 100% could use the filters, though some struggled to find the right option
• 56% preferred the dashboard view and 45% the thumbnail view for prospecting; everyone used the list view for bulk actions

After launch, the customer success team moved users over and almost everyone was migrated within a month.

• 95% of users migrated to the new platform
• 121% increase in MRR
• NPS rose from -25 to +50, a 75-point swing

What went well: Building the score and fit data as a feedback loop. The more users qualify and reject companies, the better their recommendations get. That gives users a reason to keep coming back, which was the churn problem we set out to fix.

What I'd do differently: Filters and navigation. Every tester could use the filters but some couldn't find the one they wanted, and 55% of those who liked the pipeline wanted it in its own menu. Both were next in line for iteration.

What's next: Discover became the foundation for Goava Ping, the companion mobile app, and later for Goava's AI agents, starting with the Prospecting Agent.