How to build viral and referral loops

1 The bottom line

Find the moment your users already want to show someone, and build sharing and referral into it. Virality cannot rescue a leaky product, but done well it brings in more users who also convert better.

2 Key takeaways

• Fix retention before you chase virality. Loops only add users faster, and a leaky product loses them just as fast.
  • Mark Pincus, Zynga40:08 says Zynga won with a clear mission and retention as its core metric, not spam.
  • Life360's in-app invite tests failed across the board because parents, the decision makers, are not viral Giordano Contestabile, Life36051:50.
• Find the moment users are proud to show. A natural share moment gets shared without being pushed.
  • Wrapped worked because it turned passive use into an identity-linked moment in a compressed period Jack McDermott and Tanmay Jain, ElevenLabs36:57.
  • Cal AI49:00 ran a free Thanksgiving promotion because people already post what they ate.
• Build the shareable output into the product. What users post markets the app for you.
  • Cal AI's progress photos and milestone badges go to Instagram stories and also build retention Daniel Heintzman, Sway AI30:00.
  • When Liquidity moved 100,000 subscribers to beehiiv, all of them saw the beehiiv footer the next day Tyler Denk, beehiiv20:00.
• Make every shared item invite the viewer to join. Viewers of shared content are your best pool of new users.
  • Mikey Shulman, Suno20:05 thinks roughly 90% of new people who get a shared song could be converted into making one.
  • GameChanger's roster invite acquires every guardian with no extra work from the coach Sameer Ahuja, GameChanger13:36.
• Show referrers progress, and judge the program on revenue. Referrers need feedback to keep going, and revenue shows the quality of what they bring.
  • SoFi's friend-by-friend tracker lifted referrals from 2% to 40% of new sign-ups in 3 months SoFi.
  • Headspace's Guest Pass gave a 7% rise in signups and an 8% rise in revenue Headspace.

you have to think about what your natural sharable moment in your product is

Jack McDermott and Tanmay Jain, ElevenLabs

2.1 Check the base and the sharers first

Loops multiply a product that already keeps users and has people who want to tell others.

• Fix the leak before you add loops. Mark Pincus, Zynga40:08
  • Virality-based companies are sinking speedboats that add users faster than they leak. Plug the hole or build a better boat.
  • Eight of 10 large Zynga launches became massive hits. Pincus: "It wasn't that we were good at virality."
• Find who actually spreads the word before you add share buttons. Giordano Contestabile, Life36051:50
  • About 40% of Life360 users say they found it through someone else, so the team tested in-app share buttons and gifted silver subscriptions.
  • Those tests failed across the board. Parents are the decision makers and are not viral. Their kids are, but they are not the core audience.
• Put distribution into the product from day one. Mark Pincus, Zynga1:01:25
  • Hoping for virality or word of mouth is a hope strategy. With weak discovery, go after power users who pay up front before mass consumers.
  • Morrie Schonfeld, Pingo6:40 asked which AI feature would be inherently viral and chose talking to an AI, with four colored lines as the character.
  • About 40,000 new games launched in the app store last year, and none became a top 10 hit or held a top 25 spot, per Pincus.

2.2 Find your natural share moment

Look at what users already do and are proud of, then make that easier to share.

• Look for behavior that is already spreading, and ease it. Sam Chaudhary, ClassDojo21:58
  • Each school had a vocal early-adopter teacher. ClassDojo gave them a ready-made staff presentation with a QR code at the end so colleagues could join.
  • The tactic was low tech, yet Dojo spread school by school to about nine out of 10 US schools.
  • Paul Warren, Margins35:10 did no marketing for 4 to 5 months and stayed in its category's top 200 on word of mouth from book clubs, sisters, kids and classes.
• Do not copy Spotify Wrapped. Find your own moment. Jack McDermott and Tanmay Jain, ElevenLabs36:57
  • Look for milestones users are proud of, such as the GitHub skyline, YouTube 100k or hours of agent interaction.
  • Expect to iterate several times. If people don't share it, it isn't a shareable moment.
  • A Wrapped-style recap from a coffee shop showed the pattern had jumped the shark.
• Tie the product to a shared occasion or a reward. Daniel Heintzman, Cal AI49:00
  • Cal AI gave free access at Thanksgiving because people already post what they ate on Instagram stories.
  • Drew Levin, SideShift4:38 says apps where users earn money or get something back do especially well with organic content. Examples are Settlemate, Snag and GLP-1 apps.

2.3 Build shareable output into the product

Give users something worth posting, and make the path to it better.

• Give users progress they can post. Daniel Heintzman, Cal AI28:28
  • Add streaks, milestones and before-and-after photos. People only want to download apps that people they know use.
  • Daniel Heintzman, Sway AI30:00 lists progress photo uploads, meal sharing with calories, and badges for days logged in a row. The same features build retention.
• Improve the creation journey and let users pick the channel. Mikey Shulman, Suno18:54
  • Suno users share by text, WhatsApp, social posts and slideshows of family photos. Suno did not dictate channels.
  • Every time the journey got easier or better, sharing behaviors exploded.
• Ship a free tool with no gate that leads into the app. Anada Lakra, BoldVoice44:40
  • BoldVoice made a web app that guesses where you are from in 10 seconds of speech, with no download, no gate and no trial.
  • Results show a snippet of the full app, which raises conversion. City-level results added another viral boost.
  • After an angel investor posted it on Threads, about 100,000 people in Korea tried it the next day.
• Design flows that look good on short video. An (Anup), Flame51:05
  • Flame invested in an AI loading animation and a clear home screen (decoder, replier, opener) so demos look eye-catching. Cal AI's scanning feature is a similar example.
  • A racing game's first few hundred thousand downloads came from TikTok. Then top chart rankings and friends beating each other's scores made the K factor matter more than the TikToks Rafael Kramer7:05. It reached number two overall.

2.4 Make every shared item invite new people in

The person who receives your content is a new user waiting for a clear next step.

• Make it obvious on the shared page that viewers can create. Mikey Shulman, Suno20:05
  • Most new users arrive when someone shares a song. Let them listen without logging in, and say on the page that it was made on the platform and they can create too.
  • Shulman: "The biggest one is making it obvious." He thinks roughly 90% of new people who get a shared song could be converted into making one.
• Brand what free users publish. Tyler Denk, beehiiv20:00
  • beehiiv adds its branding to the footer of free users' emails and websites unless they pay to remove it.
  • Weber Wong, FLORA39:47 added project-level sharing and saw more signups from organic sources.
• Let one decision maker pull in the whole group. Sameer Ahuja, GameChanger13:36
  • The coach enters the roster, and each guardian gets an invite to download the app and claim their athlete. GameChanger added a web link so coaches can type 20 guardian emails on a computer.
  • Shared highlight clips brought non-using relatives back, and parents began asking coaches to stream games Ahuja32:30.
  • GameChanger saw a step change in fan community size, not linear growth. The average user stays six or seven years.
• Turn video attention into installs with comment-to-DM. Casper, Payout15:16
  • Ask viewers to comment a keyword, then use ManyChat to DM them the app link. The DM notification stays until they act.
  • About 13,000 commenters on one video led to about 6,500 clicks, around 50%. Download to paid was about 6.7%.

2.5 Add referral rewards and track them

A referral tracker took SoFi from 2% to 40%

Before tracker2%

After tracker40%

Show referrers where each friend stands, reward them early, and measure what the friends are worth.

• Show each friend's progress and the reward at stake. SoFi
  • Replace "you learn it worked when the reward lands" with a tracker of each invited friend's sign-up stage. Show potential earnings throughout the flow.
  • Visibility removes doubt about whether sharing is worth it and gives referrers a reason to follow up with friends who stall.
  • Before the change, referrals were only 2% of new users. After it they were 40%, up 20x within 3 months. SoFi treated the low share as a design problem, not only an incentive problem.
• Reward referrals before the user pays, and pick a reward that fits. Daniel Heintzman, Cal AI30:40
  • Cal AI tested $10 for referring a friend, available during onboarding before the user pays.
  • In AI products where power users buy extra credits, reward invites with credits Phil Carter, Elemental Growth33:56.
  • Cal AI could not isolate the impact, because word of mouth is hard to track.
• Give referred friends a longer trial. Headspace
  • Headspace's Guest Pass gave referred users 30 days free, double the standard 14, so skeptics had time to feel the value.
  • A recommendation from someone they know lowers the barrier to start. Referral-driven trials converted to paid better than the standard flow.
• Offer invite-to-unlock with a branded share template. Nicole, Glam Up18:38
  • After a face scan, Glam Up blurred results and offered two buttons: pay, or invite three friends to unlock free.
  • The share action pastes a template with the app name, not a bare code Nicole19:28. Thousands of TikTok comments made the branded keyword rank in search.
  • The app logged about 1.6 million share requests and earned over $600,000. Many sharers ended up paying anyway, because getting three friends to onboard is slow.
• Judge the program on revenue, not only signups. Headspace
  • Headspace tracked both. Referred users converted from trial to paid at a higher rate, so the gain came from quality as well as volume.
  • Cal AI said most growth still came from influencers and paid ads, and sharing is hard to track Daniel Heintzman28:28.

2.6 Where they disagree

Which loop fits depends on who your users are and what they make.

• Referral features work for some audiences and fail for others. SoFi
  • SoFi's tracker took referrals to 40% of new sign-ups.
  • Life360's in-app invites failed because its decision makers, parents, are not viral Giordano Contestabile51:50.
• Gate the output or leave it open. Anada Lakra, BoldVoice44:40
  • BoldVoice gated nothing and let results flow freely into the app.
  • Glam Up18:38 blurred results until users paid or invited three friends, and conversion did not drop much.
• Virality as a core bet or a side effect. Morrie Schonfeld, Pingo6:40
  • Pingo's team said there is no point in building a consumer app that won't go viral.
  • Zynga's Pincus says retention, not virality, drove its hits Mark Pincus40:08.
• Loops outside your own app fit platforms and AI products. Casey Winters, SuperMe25:00
  • Let users tool call your product from Claude or Google Docs, so they get value without visiting your site. Profile and content sharing got SuperMe its first 3,000 users.
  • Tyler Denk, beehiiv47:10 uses creator case studies as a loop, since audiences of successful creators often become creators on the same platform.