Doubling response rates through service design

Redesigning the end-to-end service experience of quarterly data collection, from a 7-week manual ordeal into a system that gives founders something back.

My RoleService & Systems Designer

MethodsUser Research, Journey Mapping, Information Design

StakeholdersFounders, Stream Leads, Coaches, IT, Leadership

ImpactResponse rates: 45% → 80%+

The Story

Hundreds of startup founders filled out this check-in every quarter, and most of them dreaded it. I redesigned the whole service to make it worth their time.

District 3's quarterly check-in collected existentially important data. It fed reports to Concordia University, Quebec government agencies, federal agencies, and partner nonprofits. It directly impacted funding. But the experience of providing that data was broken for everyone involved: founders ignored emails, stream leads chased people through busy inboxes, and the data team drowned in manual processes.

Automating the old process would have preserved a bad experience at higher speed. The job was to redesign the service so people wanted to participate.

The Journey

Inherited Pain

Summer 2022

Joined D3 as Data, Systems & Automations Coordinator. Inherited the manual quarterly survey process and its 7+ week cycle.

Meeting-Hell

2022 - Early 2024

Lived through multiple survey cycles. Attended endless meetings about how broken the process was. Watched response rates crater from 90%+ pre-COVID to the mid-40s. Nothing changed because there was never time to change it.

The Catastrophe

April 2024

While studying interaction design in Paris, I accidentally deleted the entire Check Ins Podio app. Restored from backup scripts, but had to rebuild all automations from scratch. Decided: if I'm rebuilding anyway, I'm going full hog.

The Rebuild

Summer 2024

Returned from Paris and started rebuilding with intention. Designed the interconnected Podio app ecosystem and the philosophy shift from taking to giving.

Automated Emails v1

September 2024

Deployed the first version: automated email sequences, stream lead validation flows, and the human-in-the-loop design. The 7+ week manual process became a monitored system.

Nutrition Labels & Aftercare

Late 2024 - Early 2025

Added Startup Nutrition Labels, the thank-you email system, data cleaning automations, and all the 'giving back' features. Response rates climbed to low-80s.

The Handoff

Mid 2025

Trained my replacement by co-creating a parallel system for annual alumni surveys. Left D3 knowing the system would outlive me. It's now run without a hitch for over six months.

The Service Design Problem

I was designing for a multi-stakeholder service with four distinct user types, each with different needs, pain points, and touchpoints. Understanding this ecosystem was the first step.

👩‍💻 Founders

Need: Minimal friction, clear value exchange

Pain: Time-poor, saw surveys as obligation with no return

Touchpoints: Emails, form, receipt, Nutrition Label

👥 Stream Leads

Need: Quick validation, predictable timeline

Pain: Chasing founders through busy inboxes

Touchpoints: Validation Excel, email replies, final reports

🔧 IT Team

Need: Monitoring capability, manual override when needed

Pain: Manual export/import cycles, human error

Touchpoints: Validation checkpoint, schedule controls, error flags

📊 Leadership

Need: Reliable data, strong response rates for stakeholder reports

Pain: Low confidence in data, endless meetings about the process

Touchpoints: Final reports, response rate metrics

The old process served none of these users. Sixteen manual steps stretched across 7+ weeks, each handoff created a bottleneck, and founders mostly experienced the whole thing as being chased.

Founder Journey: Before & After

The most important user in this service is the founder. If they don't submit, nothing else matters. I redesigned their journey from first contact to follow-up:

Stage	Before	After
Awareness	Surprise email arrives with no context	Coaches mention upcoming check-in during sessions; founders expect it
First Contact	Generic email with blank form link	Personalized email with pre-filled form URL (startup name, quarter, known constants)
Form Experience	Confusing questions, no guidance, ~30+ minutes	Logically grouped questions, clear help text, sticky header with dates, ~15 minutes
Reminders	Increasingly terse emails citing agreement obligations	Friendly nudges with clear value proposition
Submission	Submit into void, no confirmation, no feedback	Instant PDF receipt, thank-you email with resources, discount reminders
After	Data disappears into black hole, never see it again	Personalized Nutrition Label arrives with their data visualized over time

The Trade

The old approach was all taking: "fill this out because you agreed to in your contract." I rebuilt the service around a trade. Founders give us fifteen minutes of data, and we give them back a personalized report on their own startup.

Form Design: Reducing Friction Within Constraints

Podio's webform builder is restrictive; no progressive disclosure, limited custom CSS, rigid field types. I optimized the form experience within those limits:

Information Architecture

I grouped the 25+ questions into logical categories that match founders' mental models, keeping related information together so they could batch their thinking:

1. Basic Context

Startup selection (pre-filled), quarter, website updates, company description

2. Revenue & Sales

90-day CAD metrics, geographic breakdown (Canada/US/Other), recurring revenue %

3. Customer Growth

New vs. recurring clients, market opportunities

4. Funding & Capital

Grants/subsidies, debt, equity, separated for clarity

5. Team & Employment

FT/PT breakdown, Canada/international, Concordia association

6. Legal & IP

Patents (pending/granted), NEQ, BN, headquarters

7. Support Needs

Where they need help (multi-select), specific requests (relayed to coaches)

8. Feedback & Comments

NPS, coach ratings, what's working/not working, newsworthy highlights

Friction Reduction Tactics

Pre-filled URLs

Each founder receives a unique URL with their startup name, the quarter, and known constants already populated. They skip the busywork and get straight to the new data.

Sticky Header with Key Dates

Despite Podio's CSS restrictions, I added a persistent header showing quarter dates (Q1: Jan 1 – Mar 31, etc.) so founders never have to scroll back up to check the reporting period.

Carefully Curated Microcopy

Every help text was written to reduce confusion: "No negative numbers, please." "Only share your website URL if it has changed." "Include founders in this count if they fall in this category."

Smart Optionality

Fields like NEQ/BN say "If you have previously provided this, leave it blank." Founders who've submitted before can skip what hasn't changed.

Time Expectation Setting

The intro states upfront: "It should take approximately 15 minutes to complete." Managing expectations reduces abandonment.

Why Your Submission is Important

I added a "Why This Matters" section at the top of the form, something the old form lacked entirely. Founders now understand upfront that their data:

• Remains confidential and is only viewed by the data team
• Helps summarize outcomes for stakeholder updates
• Highlights their successes within the Concordia community
• Helps secure funding and improve D3's support

Startup Nutrition Labels: Information Design

The Nutrition Labels are the centerpiece of the "giving back" philosophy: automated analysis reports that turn founders' submitted data into something they can use. The hard part was the information design.

Research: What Founders Care About

Before designing the reports, I needed to know which metrics founders track. I used two research methods:

Qualitative: Stream Lead Interviews

I interviewed stream leads who coach founders daily. They identified the five areas founders obsess over: revenue trajectory, funding runway, team growth, customer acquisition, and overall business direction.

Quantitative: Existing Data Analysis

I analyzed responses to "Where do you need the most help today?" across historical submissions. The trends confirmed the interview findings, these five areas dominated.

The Five Charts

1. Sales Revenue by Region

Stacked bars showing quarterly revenue from Canada, US, and elsewhere, with a smoothed moving average line for total sales.

Design decision: The smoothed line filters out quarter-to-quarter noise and reveals the actual trajectory, founders can see if they're trending up or plateauing.

2. Funding Sources & Timeline

Stacked bars for equity, debt, and grants per quarter, with a cumulative line showing total capital raised over time.

Design decision: Separating funding types helps founders see their capital mix. The cumulative line is what they'd show investors, "we've raised $X total."

3. Workforce Distribution & Growth

Stacked bars for FT/PT employees in Canada and internationally, with smoothed total headcount trend line.

Design decision: Team composition matters for different reasons, immigration, payroll planning, remote work strategy. Breaking it down serves multiple needs.

4. Client Acquisition & Retention

Stacked bars for new vs. repeat clients per quarter, with smoothed total trend line.

Design decision: A startup with mostly repeat clients has product-market fit; one with mostly new clients is still searching. The split makes that visible at a glance.

5. Business Focus Word Cloud

Aggregates text from "Specific Need," "Achievements," "What's Working Well," "What's Not Working Well," and "Additional Comments" across all their submissions.

Design decision: Qualitative data is hard to visualize. The word cloud surfaces recurring themes, so founders can see which topics they kept coming back to (or worrying about) over time.

Visual Design: The Nutrition Label Metaphor

The report is styled minimally in black and white, directly inspired by actual nutrition labels. The metaphor works on multiple levels:

• Familiar format: Everyone knows how to read a nutrition label
• Objective data: Just the facts, presented clearly
• Health check: A periodic assessment of your startup's "health"
• Standardized: Same format every time, easy to compare quarters

Report Structure

• Startup name and quarter
• Five charts with methodology explanations
• Footer disclaimer about data limitations
• Subtle invitation to backfill past quarters

Technical Flow

The Nutrition Label generation required careful timing due to Podio's quirks:

1. Submission triggers creation of a Nutrition Label item linked to the startup
2. System waits 3 days for Podio to "settle" (calculation fields are rate-limited and queued)
3. Automation fetches all historical Quarter ↔ Startup pairings for that startup
4. QuickChart generates visualization URLs from the aggregated data
5. WGET saves chart images to Podio fields (workaround for PDF generation)
6. PDF is compiled and emailed to cofounders

Validation

Before building this, I gauged interest in the community Slack. The response was enthusiastic. After the first round of Nutrition Labels went out, two founders followed up specifically to share positive feedback, and in subsequent user interviews, founders described the new system as "an all-around upgrade" that made them "much happier to take the time."

Designing for Failure

The happy path was the easy half. I catalogued the ways submissions could go wrong and built automations to catch, clean, or flag problems before they corrupted our data.

Automated Cleaning

• Negative numbers: Flagged automatically (you can't have -3 employees)
• Suspiciously high numbers: Flagged for manual validation (did they really raise $50M?)
• Duplicate submissions: Detected and flagged before they skew reports
• Malformed data: Cleaned or flagged based on field type

Automated Updates

• NEQ/BN numbers: Auto-update central Startup card
• Website URLs: Auto-update if changed
• Headquarters address: Auto-update central record
• Patent counts: Auto-update IP tracking
• Newsworthy highlights: Auto-forward to Communications team

Solving the "Black Hole" Problem

Previously, founders would submit their website, address, and legal numbers, and that data would sit in a spreadsheet, never updating the central database. Now, every submission automatically enriches our primary Startup records. Data flows where it's needed without manual intervention.

I was also able to run these cleaning scripts retroactively on years of historical data, significantly improving data quality across the entire dataset.

Designing for Trust: Human in the Loop

I kept humans in the loop at key moments. Full automation was technically possible; I chose against it.

This process emails hundreds of busy founders. Sending unnecessary emails damages the organization's reputation and trains people to ignore future messages. A quick IT review before triggering the email cascade is worth the extra step.

IT Validation Checkpoint

After pairings are generated, IT receives an email to skim the list before triggering the email sequence. They can adjust send dates if stream leads are slow to validate.

Stream Lead Validation

Stream leads receive Excel files with custom "do not contact" URLs. One click removes a startup from the list, catching cases where teams left but weren't updated in the system.

The two checkpoints cost the team minutes per cycle, and they prevent the expensive mistakes: emailing a founder whose startup already left, or triggering the sequence before stream leads have validated the list.

Results

Quantitative Impact

Mid-40s% → Low-80s%

Response rate increase within two quarters

1 Month → 1 Week

Stream lead validation time reduction

Hundreds of Hours

Saved per quarter across all stakeholders

Manual follow-ups required

Qualitative Improvements

• Founders: User interviews found they saw this as "an all-around upgrade" and were "much happier to take the time"
• Stream leads: Appreciated the simplicity and predictability of the new flow
• IT: Shifted from manual labor to light monitoring
• Leadership: Escaped meeting-hell; gained confidence in data quality
• Data quality: Automated cleaning + retroactive fixes improved the entire historical dataset

The Handoff

Six months ago, I left D3. Before I did, I trained my replacement by co-creating a parallel system: an annual check-in for alumni and graduated startups. Building something together was the best way to transfer knowledge of Podio's quirks and the design decisions embedded in the system.

Since then, she's run two full quarters of the quarterly system without a hitch. I wrapped up a follow-up contract with D3 last week, the system is still running exactly as designed.

Reflection

This project taught me to design the entire journey: all four user types, failure modes included. The old check-in's technology worked fine. No one had designed theexperience of participating in it.

The shift from taking to giving changed how I think about organizational processes in general. A process that asks people for their time should give something back. The Nutrition Labels cost me development hours and repaid them in response rate: founders went from being chased to submitting on their own.

The accidental database deletion helped more than it hurt. It gave me a blank canvas and permission to rebuild with intention instead of patching what existed.

Keeping humans in the loop was a design decision, and I would make it again. Where the stakes involve hundreds of founder relationships, a person should review before the system acts.

The Check In Suite has outlived my time at D3. For a service design, that is the result I care about most.

Back to works