PLG In-App Promotion: 0% → 26.7% via data-driven design

Free Trial application conversion rate.

My Impact

• Identified GMs and OPMs as the target audience and used GA4 data to back-derive the entry-point design
• Proposed 5 placement candidates and converged to 2 banners based on data
• Designed the in-app modal flow, reducing the form from external HubSpot to 4 auto-filled fields
• Aligned the team with data and a risk-buffer plan when the PM was uncertain about placement
• Achieved 100% activation across 8 targeted companies in one month, establishing a framework later reused for cases like In-Entry Audit

The Brief

This was a Growth Design project, the second phase of GoFreight's attempt at PLG (Product-Led Growth, where users discover and self-convert through the product itself). I owned the redesign of the Free Trial entry flow.

Phase 1 (Free ISF) had ended with 0 applications submitted. To kick off Phase 2, I ran a workshop with the Senior PD, Junior PM, and Senior PM, walking through Phase 1 end-to-end and brainstorming redesign directions. The hypotheses were then validated against the data.

Challenge

1. 1.High-friction content format. The "What's New" surface bundled many unrelated announcements together, so the promo rarely registered as relevant and most users skimmed past it. Among those who did click in, the 35-second video lost them at 11 seconds on average, before reaching the CTA.
2. 2.Broken conversion path. The flow exited the product to an external HubSpot form with too many fields, pushing users out of context.
3. 3.Audience mismatch. The flow targeted OP operators, who had no decision authority. People with interest couldn't act on it.

Targeting: GM and OPM as dual-role users

There's a common bias around GMs and OPMs: intuitively they're decision makers, so they shouldn't sit on operational pages. But in small-to-mid-sized forwarders, GMs and OPMs often double as OPs. They open invoices and check shipment status on HBL List themselves.

To find where they actually spend time, I pulled GA4 for the pages GMs and OPMs visit most. Three came up: HBL List, Invoice List, and the Entry interfaces, all daily working surfaces.

That grounded the placement logic for what came next: target the surfaces they pass through in their OP capacity, and avoid decision-maker-only pages like dashboards or executive reports.

CTA wording: Learn More

We considered "Try Now" and "Start Free Trial," but the next step after the click was an in-app modal form, which doesn't actually start a trial. To keep the CTA wording consistent with the next action, we chose "Learn More." The body wording of the banner was iterated together with the US-based Marketing team.

In-app modal form

Phase 1's external form had many fields and required leaving the system. This time, the in-app modal was reduced to 4 essential fields, all auto-filled with the user's logged-in context:

• Company Name
• Contact Name
• Email Address
• Phone number

Users only need to confirm and submit. After submission, a Thank You message appears, informing the user that the assigned PM will reach out within 3 business days to activate the 30-day trial. By design, submitting does not auto-activate the trial. This reflects GoFreight's sales process, which still requires PM involvement.

What changed

All 8 applying companies activated within one month, indicating real intent rather than a typical click-through. The sample is small (30 users), so the next round will scale across regions and company sizes to test whether the 26.7% holds.

Adopted elsewhere

• The In-Entry Audit case reused the HBL Card gift-card placement (one of the proposals I made but didn't develop in this case)
• The placement logic (GA4 → workflow nodes) has become a reference pattern for other in-app promotions

Reflection

The biggest takeaway from this case was the practice of designing by reading backwards from data already collected. Phase 1's records were always there; they just hadn't been treated as design inputs. Once Mixpanel, HubSpot, and GA4 data were stacked together, each bottleneck mapped to a specific design problem, and none of them required new user research to find.

Working with the team

• Data-Driven Exploration (with Senior PD): To move beyond design intuition, I took the initiative to request GA4 access from the Senior PD to analyze the behavioral patterns of our core TA (OP, OPM, and GM). This self-directed data dive allowed me to pinpoint high-traffic surfaces, transforming raw data into actionable design inputs for the promotion strategy.
• Strategic Placement & Experience Optimization (with PMs): I reassessed the existing "What's New" entry, noting its cluttered interface and low visibility. By presenting GA4 frequency data to the PM team, I advocated for relocating In-App Promotions to pages with higher relevance and traffic, ensuring the message reached users where they were already working.
• Friction Reduction via Native Integration (with PMs): I identified a critical drop-off risk in the legacy HubSpot form, which required external redirects and redundant clicks. I proposed migrating to a natively developed form, which streamlined the conversion path by auto-filling account data and maintaining a cohesive system experience, drastically reducing user friction.
• Technical Feasibility & Prioritization (with Engineering): When engineering constraints surfaced regarding Global Navigation components, I leveraged GA4 rankings to lead a scope-management discussion. We collectively decided to focus development on the top two highest-impact placements (HBL List and Invoice Entry), ensuring a high-ROI delivery within the project timeline.

What this opens for B2B forwarder SaaS

For B2B, the PLG entry can be understood as workflow-embedded design: place the action on a page where the user is already going to pause, and make it light enough that they can take it without effort.

This experience also made the longer-term implications of PLG in the freight forwarder space more visible. Forwarder SaaS has traditionally been sales-led: a customer needs a demo, contracts, and onboarding before reaching the product. A working PLG path proves that even ops-heavy B2B software can support self-serve trial, opening a low-friction entry for small-to-mid-sized customers (those who previously wouldn't qualify for a full demo cycle). It may also influence packaging logic over time, shifting from one enterprise package toward independently-trialable modules.

Open questions for the next round

• What kept 22 of the targeted users from clicking the banner; qualitative interviews would help answer this
• A/B testing the CTA wording to validate whether "Learn More" is the optimal choice
• Exploring whether auto-activation upon submission could lift the completion rate further