Product Case Study

BD4

Onboarding &
Coaching System

From setup-first to value-first, and the 30-day coaching arc that followed.

Role

PM & Product Designer

Platform

iOS · Android

Status

Shipped

“Onboarding is the first impression of the product’s judgment.”

Section 01

What is BD4

BD4 is a daily business-development engine, it tells you who to reach out to, and what to say, every day.

Know who to reach, and what to say

Daily recommendations built from your own data across multiple sources, not generic outreach.

Act in a few clicks

From recommendation to sent message in seconds. BD4 does the work; you confirm.

Handle every relationship

In-app assistance and insights keep each relationship warm, so nothing goes cold.

Build your community

Surface the right people from your own network using BD4’s internal search, no manual list-building.

BD4 home screen with a daily recommendation card, today’s top action, and a BD4-drafted message

The problem in onboarding

Splash carousel → “A few things you can do” → “You can skip any step” → Integrations → Personalizing.

5-Step Setup

A few things you can do before you start

The more context you give BD4, the smarter it gets. You can do these in order or one at a time.

Connections & Privacy

What BD4 connects to

Used only to create your BD4 Professional Profile. We don’t post or access your messages.

Your First Path

Here’s what happens next

Once setup is done, you’re ready to go, but value only arrived after all of it.

Welcome, the value promise, shown before the setup ask

The 5-step setup wall

“You can skip any step”, but skipping limits features

Connections & privacy

“Here’s what happens next”, value still gated behind setup

Breaking down the pain point

01
The Assumption · The Hypothesis

Better personalization of the user profile creates better recommendations.

Philosophy, building around the user’s world.

02
The Requirement

Collect everything before the user starts.

Build Profile→Define ICP→Connect Data→Set Preferences
03
The Experience

Users paid the full setup cost before seeing their first recommendation.

The Consequence

Completed setup

High effort · Weak first recommendation

Value not proven

Skipped setup

Low effort · No recommendation

Value not experienced

Neither path built the trust the product needed.

Section 02 · Research & Discovery

What feedback surfaced

“You asked me to set up a voice profile before I had seen a single email the product could help me send. I guessed, and then I ignored everything it generated.”

MSP user, June 2026

8+ min

Time to first value

~15%

New-user activation

~18%

D30 retention

The Insight

Setup steps that precede value are answered with low-quality inputs, the user doesn’t yet understand what the inputs are for. A voice profile set up in a vacuum.

Three structural problems

01

Setup before value

Every step asked the user to give something, profile, ICP, writing samples, before BD4 had shown it could use them.

02

Cold start is existential

The next-best-action engine needed contacts to rank and signals to learn from. An empty BD4 had no substrate for its intelligence.

03

Nothing after onboarding

V1 ended with no continuation. The product was smart, but it did not help the user become smarter at using it.

The requirement

Show value before asking for trust.

Time to first value should feel like a wow moment, not a prerequisite checklist.

Why personalization still matters

It proves BD4 understands the professional’s world. It should deepen demonstrated value, not block access to it.

The Mismatch

The Product Promise

DAILY VALUE

Immediate and useful

The Setup Experience

UPFRONT EFFORT

Heavy and trust-dependent

What users wanted after onboarding

1

Tell me what to do next for a few weeks

Don’t just show what’s available. BD4 surfaced recommendations but didn’t coach the behavior that made them better.

2

Stop when I don’t need it

Coaching scales inversely to competence. A one-contact user needs different prompts than a forty-contact user.

3

The daily list is the product

“Who do I reach out to today?” is the core job. Everything else is infrastructure for the list.

The founding debate

Founder’s Instinct

Make BD4 as personalized as possible from the first moment, collect as much user information as possible before the product starts.

“More context up front = a better first recommendation.”

The PM (me) Position

Personalization without demonstrated value is just a data-collection form. The user needs to trust the product before investing their data in it.

“Nothing skippable, deferrable, or doable another way belongs in initial onboarding.”

How do we hook users by showing how smart BD4 is before asking them to do a lot of setup?

Section 04 · The Strategic Reframe

From setup-first to value-first

Setup-first says

“Complete your profile before you can use us.”

This asks for trust.

Value-first says

“Here’s what we can do for you now, and how to make us better.”

This earns it.

The redesign did not remove personalization. It relocated it, after the product had proven itself.

Section 04 · Direction

Identifying the feature that gives the most with the least input

First 10 min

Time to first value in V1

<1 min

Target, adding one contact

Value here is directly proportional to value from completing a suggested action.

30 days

Coaching arc introduced

The natural-language add, the feature that gives the most with the least input

Solutioning direction

Make the first experience action-led, then prove value through continued coaching.

Direction: 01 Start with use, 02 Reveal the intelligence, 30 days coach toward value
Section 04 · Principles

Six principles, written before any screen

1

Value before setup

Nothing deferrable belongs in initial onboarding.

2

We do it for them

BD4 leads with suggestions; the user confirms via yes/no.

3

The honest-card rule

Never state as fact anything not given or integration-surfaced.

4

Coaching scales inversely to competence

Prompt cards fall away as the user matures.

5

Progress over calendar

Prompts fire on state, not on which day it is.

6

Reuse, do not rebuild

Reposition existing components; no net-new where one serves.

Section 05 · Architecture

Built in three parts, in sequence

Each part was a separate engineering workstream, its own FRD section, acceptance criteria, and GitHub epic.

Part 1
Splash carousel, example data, on a loop

An auto-playing intro on example data, positioning → enrichment → recommendation → actions → daily list.

Positioning, your daily business development engine
Enrichment, a prospect card with signals
Recommendation, who to contact today and why
Actions, the outreach brief
The daily-list explainer
Your first-time home screen

Part 2
Landing & first flows

The first-time home replaces an empty state with one clear path, “Try with one person,” plus Build your community.

First-time home with Try-with-one-person and Build your community
Try with one person

One contact, one complete flow

Five steps, not forty, each one earning the next. Scroll to move through the flow.

Add someone you know BD4 enriches the profile Added to your community Recommended next step Outreach message ready
  1. 1

    Add

    One contact via name or LinkedIn URL. Enrichment runs.

  2. 2

    Enrich

    BD4 shows the enriched contact from just a name.

  3. 3

    Celebrate

    “Found [name], here’s what we know.” The first wow.

  4. 4

    Recommend

    First recommendation. What to do next.

  5. 5

    Complete

    Review, tweak, send, you always have the final word.

Section 07 · Flow C Coaching
Part 3

Prompt cards & 30-day coaching

Up to 2 prompt cards from four categories, a daily selection algorithm, and a 30-day phase arc.

30 days of coaching arc, Grow, Deepen, Reward rising from day 1 to day 30, Day 12 marked

Strategy behind the 30-day arc, driven by contact count

Grow

Add more contacts

Deepen

Add context, connect

Rescue

Re-engage a lapse

Reward

Acknowledge progress

The Phase Arc, Contact Count, not Calendar Days

Unlock

0–4 contacts

Grow, almost exclusively

Reach the 5-contact unlock

Build core

5–29 contacts

Grow + Deepen, alternating

Build toward 30 contacts

Sustain

30+ contacts

Deepen + Rescue + Reward

Volume prompts stop; Grow disappears

A user who adds 30 contacts on day one sees Sustain prompts immediately. State, not calendar.

The proposal on the table

Drive a daily-action quota, push the user to complete ~5 actions every day. Keep them active by keeping them busy.

This added complexity to the coaching theory and optimized for activity, not for the outcome that actually earns trust.

My position

Spend the first 30 days helping the user build their community and context, not hitting a quota. Prompt for the relationships and signals that make recommendations sharper.

Better recommendations are the proof. “BD4 understands me” is earned by showing it can do better, not by replacing the user’s whole BD process on day one.

Solution

The first 30 days should be about building community and context so the recommendations earn trust.

Example prompt cards across stages, keep going, list is live, go automatic, sharpen context, connect calendar
Section 07 · The Coaching Engine

How the right two cards reach the home screen

Every prompt enters one system. Rules determine eligibility; checks decide what survives.

Predefined library (Grow, Deepen, Rescue, Reward) funnels through scenario rules, contact count, recency, gaps, milestone, past the checks gate to the two highest-priority survivors: add 3 contacts, reconnect with 4 relationships
Why it matters

Predefined content × state-based eligibility × checks gate.

The logic is explicit in the FRD, engineering never guesses.

Section 09 · Reflection

Three versions, one direction

V1

Setup-first

4-step wizard before any recommendation. Skippers and completers hit the same empty home.

V2

Add-first

Moved add-contact ahead of setup. Sequence mattered, but still setup-heavy, still silent after.

V3

Value-first

Example-data splash → try-with-one-person → first recommendation → 30-day coaching. Full architecture change.

Current

10 · Outcomes · Success Measures

Value in the first minute. Momentum through day 31.

The system succeeds only if early value becomes sustained behavior, and sustained behavior becomes retention.

First Session

≤ 60 sec

Time to first value

Previously: 8+ min

Week 1

≥ 70%

Unlock the daily list

Reach 5 useful contacts

Day 7

≥ 45%

D7 retention

Early value keeps users returning

Day 30 → 31

≥ 30%

D30 retention

Previously: ~18%

≥ 80%

Reach the first recommendation

Almost everyone sees value in session one

≥ 45%

New-user activation

Previously: ~15%

≥ 25%

Prompt-card action rate

Per surfaced recommendation

≥ 25%

Trial → paid conversion

A populated product proves its value

Fast activation gets users in. Thirty days of coaching gives them a reason to stay.

10 · Outcomes

What shipped

✓

V1 onboarding

Shipped, 5 splash screens, 4 personalize cards, 12-rule business summary.

✓

Redesign FRD v2

Complete, 11 sections, 6 principles, 5 configurable constants, all state machines specced.

✓

Coaching system

Complete, 4 card categories, daily selection algorithm, 30-day phase arc, suppression logic.

✓

3-part GitHub epic

Splash carousel; first-time home + try-with-one; add-many/mining; prompt cards, all FRD-referenced.

Section 11 · What This Proves

Strategy, not screens

Onboarding that collects data before delivering value is not onboarding. It is a transaction.

As a Product Thinker

The value-first reframe drove every design decision in the redesign.

As a Prioritizer

The three-part structure matched what could ship in sequence. Prioritization that ignores build order is not prioritization.

As a Builder

11 sections, 5 constants, 10 state-machine states. Could an engineer resolve any edge case without escalating? Yes.