Yuriy Timen
Growth advisor, former Global Head of Marketing and Growth at Grammarly
Where to hear and read about Yuriy Timen2 episodes and tests
SubversiveHow Grammarly’s Evolution from B2C to B2B Propelled It to a $13B Valuation
Lenny's PodcastHow to grow a subscription business | Yuriy Timen (Grammarly, Canva, Airtable)Yuriy Timen was Grammarly's first marketing hire and left in 2020 as Global Head of Marketing and Growth, and he now advises subscription companies such as Canva and Airtable on growth. The thing most worth learning from him is to pick the one growth engine your product's structure fits, give it a fair, designed test, and turn fully when the market pulls you somewhere new.
If it's a pivot, treat it like a pivot and don't try to have your cake and eat it too.
How Grammarly’s Evolution from B2C to B2B Propelled It to a $13B Valuation
Yuriy Timen: the story and the lessons
1Yuriy Timen's path from investment banking to Grammarly and advising
Timen came to marketing from investment banking and private equity. After 2008 he moved into agency performance marketing, which is where he learned the craft Grammarly later hired him for.
He joined Grammarly in 2012 for the people, not the product. Grammarly first hired him as a consulting client, and a three-month project turned into a full-time role. He calls the product then a glorified spell checker.
The team made him stay, and his lesson from the hire is "always try before you buy."
He left in 2020 to advise subscription companies on growth. Eight and a half years leading growth at Grammarly led to a full-time advising practice. He has advised about 15 companies since, including Canva, Airtable, Hims & Hers, Otter.ai and Flo Health.
Casey Winters and Elena Verna gave him their time when he was starting out.
Timen caps his client load at four to five companies at once. He works three to three and a half days a week, by choice. Past five companies, the cost of context switching becomes overwhelming.
For every four to five active clients, another 10 to 12 are being explored, and 30 or more companies are friends he talks to informally.
2How Timen helped Grammarly pivot from students to knowledge workers
Usage data showed knowledge workers were already using the student product. About 50% of the text run through the grammar engine was non-academic, even though the site said students and students only. Searches for resumes, cover letters and resignation letters pointed the same way.
Grammarly built the product to live wherever people write. Knowledge workers write mostly in the browser, so a web app added friction and reached only high-intent users. The team built a Chrome extension, Office plugins and a mobile keyboard.
The keyboard also opened app stores as a new acquisition channel.
He refused to paywall an experience too novel to explain. Videos and animated GIFs could not do the extension justice, so people had to use it. Casual workplace writers would not climb a paywall the way students with high-stakes papers would.
In prototype, "it would be a crime to pay wall that experience."
Grammarly tested going free-first and accepted a revenue dip. An A/B test of extension-first against the old pay-first web app tolerated up to about a 10% immediate revenue dip. Top of funnel exploded, which gave the team confidence to move the whole $30M a year business to freemium.
The bet was that converting even a small percent of the new users over time would more than cover the gap.
The pivot changed the promise, the reach and the business model together. The promise moved from getting better grades to better career outcomes. It followed Geoffrey Moore's whole-product idea from Crossing the Chasm.
3How Timen ran Grammarly's search and YouTube channels
Timen started with high-intent queries and expanded in concentric circles. Roughly 80% of early acquisition came from paid search and SEO on queries like grammar checker and essay checker. Volume rises as intent falls, so each ring needed a different content approach.
High prices and solid lifetime value let Grammarly feed its performance engine with these queries.
He shut down the long-tail forum when conversions fell faster than traffic. Queries under about 500 searches a month went to Grammarly Answers, a moderated forum. It brought lots of traffic and very few conversions and cost too much to maintain.
Grammarly lacked the user-generated content that made the extreme long tail cheap at Quizlet.
Reaching professionals meant moving from demand capture to demand generation. Professionals thought of grammar as a school habit and did not search for a checker. Grammarly told a story in order: a pain, proof, a way to solve it, then why Grammarly is best placed.
A line like the world's most accurate grammar checker stops working once you leave high-intent search.
Timen picked YouTube over Meta by thinking about the viewer's mindset. Meta ran mostly static ads then, while YouTube is more educational and viewers are more likely to be in a work mindset. YouTube became Grammarly's strongest channel for prosumers.
He says product-channel fit matters as much as product-market fit, because a channel off by a few degrees performs far worse.
4How Grammarly Business grew once Timen's team built proof for each function
Inbound requests for team seats were not the tip of the iceberg. Subscribers asked how to buy for their teams, so Grammarly shipped bulk licenses, seat management and basic reporting in 2018. The team expected an absolute explosion, and it did not come.
Inbound requests did not mean the full B2B product was ready.
Selling to managers worked only after going vertical. Individual-ego pitches and a generic 'minimize misunderstandings' message were hard for managers to fund. The team started with marketing, sales and customer success and asked what each is graded on.
Proof points were CSAT scores for customer success and win rate for sales, built through pilot customers. Team licenses now grow much faster than single seats.
The green spinning G spread Grammarly by word of mouth inside companies. Colleagues noticed the G during screen shares and meetings and asked what it was. The effect was impossible to measure, but the team knew it was powerful.
Users grew so attached that a missing G after a browser update felt devastating. Most word of mouth happens offline, so being hard to measure does not make it unimportant.
5What Timen would do differently at Grammarly
Timen says Grammarly was late to both knowledge workers and B2B. Users were already asking to buy seats and move accounts between emails. He says Grammarly should have gone after knowledge workers earlier and built team features like collaboration from the start.
He regrets leading from fear of losing rather than a drive to win. Grammarly was bootstrapped for its first seven years, which bred a scarcity mindset and blind spots. It was late to react to LLMs and scrambled once they became commercial products.
"I wish that we would have led from a position of trying to win as opposed to trying not to lose."
A real pivot is a full turn, not a marginal correction. Timen says a pivot means the current course was not working. If a pivot is needed, burn the boats and go all in, because survival is often at stake.
A change that leaves the current course intact means either the pivot was never needed or the response fell short of one.
Grammarly moved up the value stack, but reacted late to LLMs. Grammarly built its grammar models from scratch with a large NLP team, then moved to tone and higher-level feedback. A vision deck from about ten years earlier planned a move from revising text to composing it.
He expects value to go to application-layer apps with user trust and reach, since no application will be built on a single model. Grammarly has 30 to 40 million or more daily users.
6How Timen picks a growth engine for a subscription product
Timen reads the product's structure to choose paid, virality or SEO. Healthy economics point to paid, built-in network effects to virality, and a long tail of search terms to SEO. A free-to-paid rate of 5% or more on top of strong LTVs signals room for paid loops.
Canva's first loop was long-tail SEO on make and template keywords. Grammarly and Canva have single-player LTVs in the hundreds of dollars, against $50 to $60 for a typical $5 to $7 consumer plan.
He backs virality only when network effects exist from day one. Airtable, Whimsical and monday.com have them. Grammarly lacked them because writing is not an inherently multiplayer task, and it found engineering them an uphill battle.
"It's very hard to manufacture product network effects if they aren't there from the get-go." A beloved brand can also run referrals, as Lyka does with give-one-get-one.
Timen vets SEO with three checks before committing. Does the company have something unique for how-to results, a programmatic angle like Canva's templates, and a unique data angle like Monarch Money's spending data? Two of three puts it in good shape, then he time-boxes a test to three months.
He calls this a back-of-the-envelope framework.
He buys an outside SEO audit first and expects slow returns. Audits cost $5 to $10K and save many human hours, since the basics no longer differentiate. SEO shows results about six months out as a small trickle that compounds.
A team can spend three to six months and learn it won't pay back. He rates Ethan Smith of Graphite as the most disciplined first-principles SEO thinker he has met.
Timen focuses on one engine and sets guardrails for moving on. For YouTube, set a minimum number of impressions, two to three creative angles and a click-through range. Inside the range keep going, outside it abandon and revisit later.
Spreading one growth owner across five tactics means none gets a real test. Abandoning is a pause, so review it on a schedule against the opportunity cost of other bets.
He says a channel that failed an unfair test has not been tested. When companies say YouTube did not work, the test was often never designed to succeed. Wrongly concluding a channel is dead is as costly as a failed test.
"You prematurely or erroneously concluded that it doesn't work and it's remarkable how often you find that to be the case."
Timen tells small companies to delay diversifying. Growth is usually 80/20, with one strategy working overwhelmingly well. Early teams should lean harder into it, as Grammarly did with performance marketing and Canva with SEO.
Companies at $50 million or more in revenue that are 90% reliant on one channel do need to carve out resources to explore others.
7Timen's rules for paid media and measurement
Timen says paid media must now pay back in six months or less. A 12 month payback used to be decent. With funding tight, spend that can't be attributed gets cut, and buying users at an LTV to CAC of 1 to 1 is now a no-no.
Companies with strong unit economics, like Grammarly and Canva, gain as weaker rivals pull back and ad costs fall. He stresses the shift from growth to survival is his own extrapolation.
He uses paid for fast learning as well as returns. Paid is a quick way to learn messaging, positioning and designs, such as A/B testing headlines in Google search. If funnel conversion, retention and LTV are weak, he spends only a small test budget.
A common failure is $100k a month with no clear goal and no experiments.
He proves causation with incrementality tests, not click attribution. Click and cookie attribution only shows correlation. Only controlled experiments show cause, though a clean test means turning off a channel and losing sales today.
He points to Measured and Incrmntal for incrementality testing. "The only way to determine causality is through real controlled experiments."
Media mix modeling pays off above $100,000 a month across three channels. The method dates to 1950s ad budgeting and is back because offline channels are common again and online tracking has weakened. Recast is bringing it into the digital world.
He hasn't yet used Recast with his own companies.
Timen warns against both too little and too much attribution work. Under-investors keep spending across many channels because growth coincides with spend and they fear breaking it. Over-investors chase scientific purity the payoff may not justify.
He tells brands not to dismiss TikTok or offline channels. TikTok reaches far beyond teenagers and is still unsaturated with advertisers. One brand with a core demo of married earners over 40 above $200k found its audience there.
Offline attribution has improved while online has worsened. One startup's out-of-home campaign returned 10 to 1 on the deals it produced, and podcasts perform for many brands.
8Timen's funnel benchmarks and onboarding advice
He wants 5 to 7 percent free to paid before scaling paid. Under 5% won't work long term for a company aiming to stay independent. Aim above 5 and ideally above 7%.
Strong LTVs, conversion and retention are what make paid a profitable channel rather than only a learning tool.
Timen benchmarks landing page to free signup at 20 to 35 percent. That range applies to prosumer freemium products like Grammarly, Canva and Whimsical. A small segment with strong fit can reach 40 to 50%, and broader audiences settle near 25 to 30%.
He calls onboarding almost always a big opportunity. Robust editors like Airtable, Canva and Whimsical leave new users lost. Early companies can lift activation 2 to 4x, and later stages 20 to 30%.
His fix is to collect relevant information up front and tailor the first screen so users don't face the whole product at once.
Timen invests in customer research early to give the team focus. Interviews, surveys and panels almost always pay off, especially from seed to Series B. A proper research push can galvanize a team.
"The way it can galvanize the team and give them focus and clarity and purpose is remarkable."
How Grammarly’s Evolution from B2C to B2B Propelled It to a $13B Valuation
Grammarly's first marketing hire explains how the company moved from a paid web app for students to a free-first browser extension for knowledge workers, then to Grammarly Business for teams. The lessons cover SEO, YouTube, premium conversion tests, B2B proof points and word of mouth, from a company now valued at $13B.
Timen came to marketing from investment banking and private equity. After 2008 he moved into agency performance marketing, which is where he learned the craft Grammarly later hired him for.
He joined Grammarly in 2012 for the people, not the product. Grammarly first hired him as a consulting client, and a three-month project turned into a full-time role. He calls the product then a glorified spell checker.
The team made him stay, and his lesson from the hire is "always try before you buy."
Usage data showed knowledge workers were already using the student product. About 50% of the text run through the grammar engine was non-academic, even though the site said students and students only. Searches for resumes, cover letters and resignation letters pointed the same way.
Grammarly built the product to live wherever people write. Knowledge workers write mostly in the browser, so a web app added friction and reached only high-intent users. The team built a Chrome extension, Office plugins and a mobile keyboard.
The keyboard also opened app stores as a new acquisition channel.
He refused to paywall an experience too novel to explain. Videos and animated GIFs could not do the extension justice, so people had to use it. Casual workplace writers would not climb a paywall the way students with high-stakes papers would.
In prototype, "it would be a crime to pay wall that experience."
Grammarly tested going free-first and accepted a revenue dip. An A/B test of extension-first against the old pay-first web app tolerated up to about a 10% immediate revenue dip. Top of funnel exploded, which gave the team confidence to move the whole $30M a year business to freemium.
The bet was that converting even a small percent of the new users over time would more than cover the gap.
The pivot changed the promise, the reach and the business model together. The promise moved from getting better grades to better career outcomes. It followed Geoffrey Moore's whole-product idea from Crossing the Chasm.
Timen started with high-intent queries and expanded in concentric circles. Roughly 80% of early acquisition came from paid search and SEO on queries like grammar checker and essay checker. Volume rises as intent falls, so each ring needed a different content approach.
High prices and solid lifetime value let Grammarly feed its performance engine with these queries.
He shut down the long-tail forum when conversions fell faster than traffic. Queries under about 500 searches a month went to Grammarly Answers, a moderated forum. It brought lots of traffic and very few conversions and cost too much to maintain.
Grammarly lacked the user-generated content that made the extreme long tail cheap at Quizlet.
Reaching professionals meant moving from demand capture to demand generation. Professionals thought of grammar as a school habit and did not search for a checker. Grammarly told a story in order: a pain, proof, a way to solve it, then why Grammarly is best placed.
A line like the world's most accurate grammar checker stops working once you leave high-intent search.
Timen picked YouTube over Meta by thinking about the viewer's mindset. Meta ran mostly static ads then, while YouTube is more educational and viewers are more likely to be in a work mindset. YouTube became Grammarly's strongest channel for prosumers.
He says product-channel fit matters as much as product-market fit, because a channel off by a few degrees performs far worse.
Inbound requests for team seats were not the tip of the iceberg. Subscribers asked how to buy for their teams, so Grammarly shipped bulk licenses, seat management and basic reporting in 2018. The team expected an absolute explosion, and it did not come.
Inbound requests did not mean the full B2B product was ready.
Selling to managers worked only after going vertical. Individual-ego pitches and a generic 'minimize misunderstandings' message were hard for managers to fund. The team started with marketing, sales and customer success and asked what each is graded on.
Proof points were CSAT scores for customer success and win rate for sales, built through pilot customers. Team licenses now grow much faster than single seats.
The green spinning G spread Grammarly by word of mouth inside companies. Colleagues noticed the G during screen shares and meetings and asked what it was. The effect was impossible to measure, but the team knew it was powerful.
Users grew so attached that a missing G after a browser update felt devastating. Most word of mouth happens offline, so being hard to measure does not make it unimportant.
Timen says Grammarly was late to both knowledge workers and B2B. Users were already asking to buy seats and move accounts between emails. He says Grammarly should have gone after knowledge workers earlier and built team features like collaboration from the start.
He regrets leading from fear of losing rather than a drive to win. Grammarly was bootstrapped for its first seven years, which bred a scarcity mindset and blind spots. It was late to react to LLMs and scrambled once they became commercial products.
"I wish that we would have led from a position of trying to win as opposed to trying not to lose."
A real pivot is a full turn, not a marginal correction. Timen says a pivot means the current course was not working. If a pivot is needed, burn the boats and go all in, because survival is often at stake.
A change that leaves the current course intact means either the pivot was never needed or the response fell short of one.
Grammarly moved up the value stack, but reacted late to LLMs. Grammarly built its grammar models from scratch with a large NLP team, then moved to tone and higher-level feedback. A vision deck from about ten years earlier planned a move from revising text to composing it.
He expects value to go to application-layer apps with user trust and reach, since no application will be built on a single model. Grammarly has 30 to 40 million or more daily users.
How to grow a subscription business | Yuriy Timen (Grammarly, Canva, Airtable)
Yuriy Timen explains how to pick a growth engine for a subscription product (virality, SEO or paid), how to judge when a channel has had a fair test, and how funding conditions change which channels make sense. He gives concrete benchmarks such as 5 to 7 percent free to paid conversion, 2 to 4x activation gains from onboarding, and sub-six-month paid payback.
He left in 2020 to advise subscription companies on growth. Eight and a half years leading growth at Grammarly led to a full-time advising practice. He has advised about 15 companies since, including Canva, Airtable, Hims & Hers, Otter.ai and Flo Health.
Casey Winters and Elena Verna gave him their time when he was starting out.
Timen caps his client load at four to five companies at once. He works three to three and a half days a week, by choice. Past five companies, the cost of context switching becomes overwhelming.
For every four to five active clients, another 10 to 12 are being explored, and 30 or more companies are friends he talks to informally.
Timen reads the product's structure to choose paid, virality or SEO. Healthy economics point to paid, built-in network effects to virality, and a long tail of search terms to SEO. A free-to-paid rate of 5% or more on top of strong LTVs signals room for paid loops.
Canva's first loop was long-tail SEO on make and template keywords. Grammarly and Canva have single-player LTVs in the hundreds of dollars, against $50 to $60 for a typical $5 to $7 consumer plan.
He backs virality only when network effects exist from day one. Airtable, Whimsical and monday.com have them. Grammarly lacked them because writing is not an inherently multiplayer task, and it found engineering them an uphill battle.
"It's very hard to manufacture product network effects if they aren't there from the get-go." A beloved brand can also run referrals, as Lyka does with give-one-get-one.
Timen vets SEO with three checks before committing. Does the company have something unique for how-to results, a programmatic angle like Canva's templates, and a unique data angle like Monarch Money's spending data? Two of three puts it in good shape, then he time-boxes a test to three months.
He calls this a back-of-the-envelope framework.
He buys an outside SEO audit first and expects slow returns. Audits cost $5 to $10K and save many human hours, since the basics no longer differentiate. SEO shows results about six months out as a small trickle that compounds.
A team can spend three to six months and learn it won't pay back. He rates Ethan Smith of Graphite as the most disciplined first-principles SEO thinker he has met.
Timen focuses on one engine and sets guardrails for moving on. For YouTube, set a minimum number of impressions, two to three creative angles and a click-through range. Inside the range keep going, outside it abandon and revisit later.
Spreading one growth owner across five tactics means none gets a real test. Abandoning is a pause, so review it on a schedule against the opportunity cost of other bets.
He says a channel that failed an unfair test has not been tested. When companies say YouTube did not work, the test was often never designed to succeed. Wrongly concluding a channel is dead is as costly as a failed test.
"You prematurely or erroneously concluded that it doesn't work and it's remarkable how often you find that to be the case."
Timen tells small companies to delay diversifying. Growth is usually 80/20, with one strategy working overwhelmingly well. Early teams should lean harder into it, as Grammarly did with performance marketing and Canva with SEO.
Companies at $50 million or more in revenue that are 90% reliant on one channel do need to carve out resources to explore others.
Timen says paid media must now pay back in six months or less. A 12 month payback used to be decent. With funding tight, spend that can't be attributed gets cut, and buying users at an LTV to CAC of 1 to 1 is now a no-no.
Companies with strong unit economics, like Grammarly and Canva, gain as weaker rivals pull back and ad costs fall. He stresses the shift from growth to survival is his own extrapolation.
He uses paid for fast learning as well as returns. Paid is a quick way to learn messaging, positioning and designs, such as A/B testing headlines in Google search. If funnel conversion, retention and LTV are weak, he spends only a small test budget.
A common failure is $100k a month with no clear goal and no experiments.
He proves causation with incrementality tests, not click attribution. Click and cookie attribution only shows correlation. Only controlled experiments show cause, though a clean test means turning off a channel and losing sales today.
He points to Measured and Incrmntal for incrementality testing. "The only way to determine causality is through real controlled experiments."
Media mix modeling pays off above $100,000 a month across three channels. The method dates to 1950s ad budgeting and is back because offline channels are common again and online tracking has weakened. Recast is bringing it into the digital world.
He hasn't yet used Recast with his own companies.
Timen warns against both too little and too much attribution work. Under-investors keep spending across many channels because growth coincides with spend and they fear breaking it. Over-investors chase scientific purity the payoff may not justify.
He tells brands not to dismiss TikTok or offline channels. TikTok reaches far beyond teenagers and is still unsaturated with advertisers. One brand with a core demo of married earners over 40 above $200k found its audience there.
Offline attribution has improved while online has worsened. One startup's out-of-home campaign returned 10 to 1 on the deals it produced, and podcasts perform for many brands.
He wants 5 to 7 percent free to paid before scaling paid. Under 5% won't work long term for a company aiming to stay independent. Aim above 5 and ideally above 7%.
Strong LTVs, conversion and retention are what make paid a profitable channel rather than only a learning tool.
Timen benchmarks landing page to free signup at 20 to 35 percent. That range applies to prosumer freemium products like Grammarly, Canva and Whimsical. A small segment with strong fit can reach 40 to 50%, and broader audiences settle near 25 to 30%.
He calls onboarding almost always a big opportunity. Robust editors like Airtable, Canva and Whimsical leave new users lost. Early companies can lift activation 2 to 4x, and later stages 20 to 30%.
His fix is to collect relevant information up front and tailor the first screen so users don't face the whole product at once.
Timen invests in customer research early to give the team focus. Interviews, surveys and panels almost always pay off, especially from seed to Series B. A proper research push can galvanize a team.
"The way it can galvanize the team and give them focus and clarity and purpose is remarkable."