The Secret to Running Growth Sprints That Actually Work — Learn Faster, Fail Smarter

‍

vector research partners ( aka V4RP )

‍

The Secret to Running Growth Sprints That Actually Work — Learn Faster, Fail Smarter

"Every growth experiment isn’t just about lowering your CPA, getting a few new logos, or boosting your ROAS before the next fundraise. At its core, it’s a learning engine. The question you’re always chasing: How are we going to grow this business?"

Why Growth Sprints Matter

This chapter is adapted from Matt Lerner, co-founder of SYSTM, who helps startups find—and pull—their biggest growth levers. He’s run B2B growth teams at PayPal, invested at 500 Startups, and knows the pain of watching founders spin their wheels on incremental tactics. Today, he’s sharing insights from his book “Growth Levers and How to Find Them” for impatient founders and their teams.

Here’s the truth: every great startup made tons of wrong turns before hitting hockey-stick growth. The more wrong turns you can take fast, the faster you get to your big win.

Think of it this way: imagine you knew it would take exactly 1,000 failed experiments before experiment #1,001 made you a decacorn. Would you play it safe or go full chaos mode? You’d run experiments like a mad scientist. Ten per week, per person. You’d hire engineers, data nerds, maybe even actual scientists. Everyone would obsess over hypothesis design, statistical significance, and avoiding type I and II errors.

Reality check: it’s not 1,000. You don’t know the number, and it’s never that neat. But the lesson is the same: early-stage startups need to learn fast. Really fast.

Stop Wasting Energy on Small Stuff

Most founders overcommit to their favorite ideas, testing at the margins, hoping incremental tweaks will move the needle. That rarely works.

Airbnb, PayPal, Canva, Facebook, LinkedIn—all found growth by identifying a few massive levers, not by cranking small knobs harder.

Big levers aren’t random. They’re physics. Startups with limited cash and small teams can’t throw spaghetti at the wall like big companies. Every action must have outsized potential.

Finding these levers is life or death for a startup. And the key? Learning faster than anyone else.

Ask yourself:

• Are we learning something new about our customers every week?
• Do failed experiments spark real conversation and insight?
• Is each mistake making us smarter, not just frustrated?

Before building a growth machine, you must build a learning machine.

The Three Steps to Finding Growth Levers

1. Understand your customers using Jobs-to-Be-Done.
2. Map your growth model to pinpoint leverage points.
3. Rapidly experiment to see what works.

Here, we’ll focus on step 3: rapid experimentation. If your experiments don’t teach you anything, nothing magically changes. Big levers win, small tweaks waste time.

The Growth Sprint Framework

Growth sprints are like product sprints on steroids for the whole company—marketing, ops, CS, and product. Here’s what they accomplish:

1. Align everyone around growth.
2. Prioritize high-upside, high-impact work.
3. Test ideas quickly to see what actually works.
4. Share learnings to accelerate team-wide learning.

Every sprint is a learning loop. Success isn’t measured in wins—it’s measured in insight.

Step 1: Gather Your Inputs

Big levers aren’t obvious. They come from unique insights about your customers.

• Customer interviews
• Usage data
• Past experiments

Example wins from this approach:

• Popsa changed “Fast easy photobooks” → “Photobooks in 5 minutes” → 4x installs.
• Rebank published “Finance for Founders” → 22x growth in 18 months.
• FATMAP built 100,000 SEO-optimized pages → 80k qualified leads/month.
• Cronofy customized homepage demo options → 5x conversion rate.

Step 2: Ideation and Scoring

Your team’s backlog will be messy. Quickly triage ideas:

1. Key Driver: Which part of the business does this move? Traffic? Retention? Conversion?
2. Impact: How big could the win be (1–5 scale)?
3. Effort: How much work/money/time does this take (1–5 scale)?

Then, prioritize: high-impact, low-effort ideas first, but don’t ignore high-effort ones.

Tip: Identify your rate-limiting step. If onboarding churn is killing growth, don’t bother buying traffic.

Step 3: Select Experiments

Pick 3–5 experiments per week (more if resources allow). Ask:

• Which could have the biggest impact?
• Will this teach us something new?
• Is there an easier way to test the assumption?
• Do we have the resources to execute?

Step 4: Run Experiments

Hypothesize before you act. Five minutes to write a hypothesis saves hours of wasted effort.

Hypothesis format:

• Risky Assumption: We believe ______.
• Experiment: To test this, we will ______.
• Prediction: We predict ______ (metric) will move ______ in ______ direction.
• Business Effect: If correct, we’ll ______.

Avoid circular hypotheses like: “We believe lowering price 10% will get more signups. Let’s lower the price 10%.” That tells you nothing about your customer.

Instead, root your hypothesis in real customer insights:

• Users aren’t signing up because they don’t understand the product → test clearer messaging.
• Users aren’t ready to demo → offer a downloadable guide.
• Users don’t trust the guide → commission an independent expert.

Test big ideas cheaply with Minimum Viable Tests (MVTs):

• Landing page → test interest before building the product
• False door → add a button to see if users care
• Manual concierge → deliver the service by hand first

Step 5: Learn Fast

Fail fast, learn faster. After each experiment, review:

• What surprised us?
• What did we learn about customers, proposition, assumptions, or execution?
• Any follow-ups? Should we scale or kill it?

Debug failed experiments by isolating variables: targeting, attention, readability, trust, resonance, or cost/benefit.

Mindset Is Everything

The fastest-learning teams aren’t always the smartest—they’re the most honest with themselves. Leaders set the tone:

• Be curious, not defensive.
• Treat failures as data, not disasters.
• Celebrate learning over vanity metrics.

Growth doesn’t start with ideas. It starts with learning.

TL;DR: Build a learning machine first, run focused growth sprints, test big ideas in small experiments, and iterate faster than anyone else. Your next decacorn move is probably just one bold experiment away.

‍