Vaibhav Sahgal on Why Growth Teams Should Be Renamed Value Connection

Growth Is Value Connection, Not Value Creation

Most companies get this wrong from day one. They hire a "head of growth" and expect magic. But Vaibhav Sahgal, who spent 3 years leading growth at Reddit before becoming VP of Consumer Product, has a different framework.

Growth teams don't create value. They connect it. A push notification from a subreddit you care about? That's value connection. You didn't create the content or the community. You just connected it to the right person at the right time.

Growth really should be renamed to something like value creation connection. And while value connection isn't exactly a very sexy term, I do think it defines what growth teams do way better than the term growth does.

— Vaibhav Sahgal

The distinction matters because it changes who you hire and when. Value creation lives in core product. They build chat features or new content formats. Growth optimizes onboarding, experiments with channels, and ensures existing value reaches users. Both grow the product. But they're fundamentally different disciplines.

Don't Hire a Head of Growth Until You Have Product-Market Fit

Sahgal is blunt about timing. If you're pre-product-market fit, hiring a head of growth is premature. You're still figuring out what value to create. Optimizing distribution before you have something worth distributing is a waste.

That doesn't mean ignoring growth strategy entirely. You need to be intentional about your distribution model, growth loops, and network effects from day one. But execution can wait.

When you do hire, go senior. Not a VP who's lost touch with execution, but someone who's been through the early chaos. They need to analyze data, write specs, hire a team, and ship—all at once. Junior growth hires can't handle that cognitive load.

You need someone who can analyze your data. You need someone who can form a really quick strategy while they are actually writing specs and operating. They also need to start hiring, building a team, scaling, and doing all sorts of hats. Not something that I think a junior person really would be able to do.

— Vaibhav Sahgal

Sahgal recommends one interview question to test execution ability: "How would you think about driving growth for [your company]?" The right candidate won't pitch flashy ideas. They'll walk you through a process: funnel analysis, prioritization, goal-setting, roadmap formation, quarterly reassessment. Thirty minutes. End to end. No fluff.

Set Guardrails or You'll Burn Every Channel You Touch

At Zynga, Sahgal learned this the hard way. When Facebook opened notifications to developers, Zynga sent hundreds of push notifications per user per day. They spammed so aggressively that Facebook changed the entire platform to stop them.

When Sahgal joined Reddit, he built guardrails. The rule: you can only send push notifications if they hit a 10% CTR minimum. Below that threshold? It's spam. Fix targeting and relevance before you send more volume.

We said you're only allowed to send push notifications if it hits a minimum CTR and if it doesn't hit that CTR, it's spam. So unless you're hitting 10% CTR, you cannot send more PMs. You have to first fix your CTR, which means you need to make it more valuable for the users.

— Vaibhav Sahgal

Channel diversification matters for another reason: personalization. Some users prefer push. Others want email. Tolerance varies wildly. You need multiple channels so you can tailor frequency and format to each user.

Know Your User So Well You Can Picture Them

Sahgal spent 4 years working on Words with Friends at Zynga. The users weren't him. They were 50 to 60-year-old women in the Midwest who saw the game as "Fitbit for the brain." That insight changed everything.

Early on, he tried to inject his own interests—sports integrations, features he'd use. None of it worked. What did work: putting a picture of "Mary from Ohio" in the office. Mary had grandkids. Mary wanted brain exercise. Every product decision ran through: Would Mary want this?

How did he get to know Mary? He talked to her. Every day. He found real users in the product and asked about their lives, why they started playing, why they kept playing years later. Deep conversations. Real friendships.

Most founders skip this. They rely on demographics or surface-level research. Sahgal went deeper, and it unlocked product decisions that no amount of A/B testing could reveal on its own.

Your First Growth Hire Should Ship in Week Two

When Sahgal interviews candidates and asks about their first 60 days, the worst answer is: "A lot of listening and learning and talking to people and forming relationships."

The right candidate wants to ship in week two. They're itching to execute. In the first month, they should analyze your data, draft a one-page strategy, and start working with engineers to launch experiments.

Red flags: lots of talking, lots of meetings, no action. Green lights: rapid execution, clear prioritization, and an obsession with getting something live.

Sahgal made mistakes building his team at Reddit. He didn't hire fast enough—he should have spent 30 to 40% of his time recruiting. He didn't fire fast enough—3 to 6 months is enough to know if a senior hire will work. And he didn't build relationships across the company early, which created an "us versus them" dynamic that took months to repair.

When onboarding your first growth hire, give them clean data access, a couple of full-time engineers, and get out of the way. If you're relying on them to fix your data infrastructure, you're not setting them up to succeed. Their job is strategy and execution, not becoming your head of data.

One final lesson from Mark Pincus at Zynga: before investing hundreds of engineering hours in a feature, test the idea first. Sahgal's team built Fast Play, a faster version of Words with Friends, based on a loose hypothesis. It took hundreds of hours. Only 5% of DAU used it. Pincus asked the obvious question: Why didn't you put up a dialog asking users if they wanted a faster board? It would have taken hours, not hundreds. The insight stuck.