Sri Batchu on Why Two-Thirds of Your Growth Projects Will Fail

When Zillow Declared War, Opendoor Refused to Flinch

In 2017, Zillow entered the iBuying market and immediately started outbidding Opendoor on homes. The entire growth team faced a brutal choice: price match to protect volume, or stick to your pricing model and watch leads evaporate.

Opendoor didn't blink. They refused to negotiate on price with a single seller. Sri Batchu watched the decision play out in real time as a early employee at the company.

Our biggest asset and competitive moat is the ability to price accurately, and we're not going to compromise that for short-term growth.

— Sri Batchu

The payoff came years later when Zillow lost billions and exited the market before the crash. The lesson wasn't about being right—it was about arriving at conviction through first principles, then holding the line when it hurt.

Expect 70% of Your Experiments to Fail (And Build for It)

Most growth leaders won't say this out loud: across Opendoor, Instacart, and Ramp, Sri has seen a consistent 30% success rate on growth experiments. Two-thirds fail. Always.

The mistake isn't the failure rate—it's failing slowly. At Ramp, the growth team still runs on 2-week sprint cadences, regardless of team size. They measure time in days since founding, not quarters. Every all-hands, every board meeting: "Day X since founding."

What's important is that you fail quickly and conclusively. Startups should have time periods of planning that are much, much shorter on a two-week sprint cadence. So you really reduce the cycle time to output.

— Sri Batchu

The worst failures aren't the experiments that don't move metrics. They're the ones that grow the wrong way. At Opendoor, Sri's team launched a program to bid on already-listed homes. Volume spiked. So did losses. Sellers listing homes that weren't selling had asymmetric information—they knew something Opendoor's diligence couldn't catch. The adverse selection killed unit economics. They shut it down after months of bleeding.

The fix: pair every volume goal with an efficiency metric. Track profitability per program from day one, especially when you're departing from core strategy.

Simplification Beats Execution (And It's Harder Than It Looks)

Instacart is a four-sided marketplace: shoppers, buyers, advertisers, retailers. Hundreds of potential growth levers. Most executives faced with two high-impact initiatives say "let's do both." Good ones pick one.

Apoorva Mehta, Instacart's founder, would say neither.

The best ones will just say, actually, we're doing neither because they don't fit in our overall strategy and are distracting.

— Sri Batchu

Sri and a few leaders at Instacart called themselves the "VPs of No." Their job wasn't to greenlight projects—it was to kill distractions masquerading as opportunities.

When the growth team at Instacart hit 300+ people, they faced a new problem: micro-teams working on narrow problems (search speed, app load time) couldn't connect their work to company goals. How much does shaving 4 seconds off load time matter?

The solution: translation factors. If the load time team reduced latency by one second, that increased order propensity by 1.2%, which translated to 100,000 monthly active users. Every team got a translation factor. Every project could be cross-prioritized in the same currency.

North Star Metrics Are a Trap (Unless You Add Constraints)

Most growth teams pick a volume metric—MQLs, PQLs, signups—and optimize until something breaks. The problem: humans are excellent at gaming the metrics they're given, intentionally or not.

If your North Star is qualified leads, your team will deliver small, low-quality leads that never convert to profit. If it's contribution profit, the feedback loop is too slow and too far from the team's actions to matter.

Sri's framework: pick something in the middle. For B2B, he uses SQL pipeline dollars. It controls for lead quality and size, but it's close enough to the team's work that they can move it in real time.

Then add the second metric: efficiency. Volume plus ROI. One without the other is a ticking time bomb.

Hire for Intellectual Curiosity, Not Domain Expertise

Growth is too new a discipline to require someone who's "been a head of growth before." The field is small. The role varies wildly company to company. Sri looks for T-shaped generalists with one deep spike—product, analytics, or a marketing channel—and the ability to learn the rest.

His favorite interview question: What's something you're bad at that you enjoy doing?

It's not about hobbies. It's about whether someone is comfortable in the learning zone or only operates where they already have expertise. In a cross-disciplinary field like growth, you need people willing to go where they don't have answers.

Red flag: defensiveness during the case study presentation. If a candidate can't take feedback on a problem they worked on for a day in a space they don't know, they won't survive the 70% failure rate.

Green flag: humility. The best candidates walk into the panel, present their work, and say "here's where I'm uncertain" without flinching.