先选对增长引擎：Grammarly 前增长负责人谈病毒性、SEO 与付费

Timen's sweet spot is subscription products, and not only consumer ones. He also works with B2B companies that lean on consumer-style, self-serve growth loops. He has spent about two and a half years advising since leaving Grammarly, working with companies such as Canva, Airtable, Hims and Hers, Otter and Flow Health, and he keeps to four or five at a time because of the cost of context switching.

He sorts subscription businesses into rough archetypes. Companies with strong consumer lifetime value, like Grammarly and Canva, can lean into paid acquisition. Timen says typical $5 to $7 a month consumer subscriptions top out at about $50 to $60 of LTV, while the strong ones run into the hundreds of dollars. They tend to attract prosumers who use the product for work and will pay $120 to $130 a year.

A second archetype has real network effects, which opens up referral and viral loops. A third is the company that can win long-tail SEO. Canva is his example: its early growth came from owning "make" and "template" searches for almost any design project, from wedding invitations onward. Not every product suits that, but he looks for it early because it can build an enormous moat.

The first thing Timen checks is whether inherent product network effects exist, and in his experience they are there from inception or not at all. Airbnb's marketplace, collaboration tools such as Airtable, monday.com and Whimsical, and Dropbox's file sharing all have them. Grammarly does not, because writing and communication is not an inherently multiplayer task, and engineering that later is an uphill battle.

Referral loops can also work without network effects when the brand is beloved. He cites Laika, an Australian fresh dog food subscription he invested in, which gets meaningful results from a give-one-get-one referral program. His definition of a network effect is the classic one: each user gets more value as more users join, though what counts is the right users, such as the people on your team or the buyers and sellers in your market.

SEO has a longer payoff than paid. Timen puts the earliest results at around six months, and says that historically a company was probably Series B before it could afford such a medium-to-long-term bet. He sees that shifting, with strategies once reserved for Series B companies trickling down to Series A as teams look to diversify away from paid.

His audit checklist has three items. Do you have a unique editorial angle on what people search for, a unique programmatic angle (Canva's templates, or Zillow and Redfin in real estate), or a unique data angle? He mentions Monarch Money, a personal finance product whose users connect accounts, as a company with data that might become an organic search experience. If you tick two of the three, you may be in good shape, and you should time-box the experiment to roughly three months.

On who should do the work, Timen says SEO is a specialist skill. The basics are widely known: the best content wins, avoid shady backlinking and keep pages crawlable. The winners are decided in the nuances and in keeping up with algorithm changes, so he recommends an outside audit, which he says costs perhaps $5,000 to $10,000 and would take many human hours to assemble yourself. He singles out Ethan Smith of Graphite as one of the most disciplined first-principles SEO thinkers he knows.

Timen sees paid as the growth bucket most affected by market turbulence. He points to Meta's and Snap's advertising revenue as signs of a global contraction in paid media budgets. The definition of efficient has also moved: for a consumer subscription company, a 12-month payback was once decent, and now the sentiment is that you should pay back media spend in six months or less. With two-plus years of runway now expected, anything that cannot be attributed is at risk of being cut.

He thinks the contraction will create winners. Companies with strong cash positions, unit economics and payback periods, such as Grammarly and Canva, face less competition on ad platforms and lower CPMs as less efficient rivals pull back. Meanwhile heads of growth are investing in attribution stacks, and he expects budgets to return once the market settles into a new normal where attribution is better than today but never as good as it once was.

Two measurement approaches are coming back. Media mix modeling, a pre-digital method for allocating budget across channels, is returning because offline channels are common and online tracking has weakened since iOS 14; he points to Recast and suggests it fits spend above roughly $100,000 a month across three or more channels. Click-based attribution never proved causality, he says, and only randomized incrementality tests can, which is why he is excited about Measured and Incrementality.

Asked whether to pursue one channel or all three, Timen recommends focus paired with rapid iteration. Going all-in on one thing wastes valuable time if it is the wrong thing, while spreading one person across five tactics means none gets a fair test. He stresses that the only thing worse than a tactic that fails is one you wrongly dismissed without a proper attempt, as with companies that say YouTube does not work but whose efforts were never set up to succeed.

To know when to move on, he suggests setting objective guardrails where benchmarks exist. For YouTube, that might mean a minimum number of impressions, two or three creative angles and a target click-through range. Abandoning a channel does not mean never revisiting it, and teams should schedule periodic reviews that weigh sunk cost and opportunity cost. For less measurable tactics he says more judgment is needed.

The 80/20 pattern holds, even at companies that look diversified from outside. At Grammarly it was performance marketing, and at Canva it was SEO. Timen says some early-stage teams start talking about diversification when one tactic is driving 80% of a still-small user base, and he tells them it is too early. Some later-stage companies with $50 million or more in ARR depend on one channel for more than 90% of acquisition, and for them it is a blind spot.

Timen describes the previous 18 months or so as a world of abundant funding, where companies coming off $15 to $25 million Series As expected to raise again within a year at much higher valuations. In that setting, paid was almost always the answer because it delivered quick returns, and SEO was seen as something for grown-up companies. Peers and investors reinforced the message of growth at any cost.

He says investors now stress survival: extend runway, cut burn and treat growth as secondary. Paid that acquires users at an LTV-to-cost ratio of one to one is a no-go. Once burn is under control and teams have 18 to 24 months of runway, the urgency of six-month growth disappears and SEO looks more lucrative. He flags that this is his own extrapolation, not something investors have told him.

He still sees a case for paid, for two reasons: it can deliver returns at efficient unit economics, and it gives the fastest learnings on messaging, positioning and features, for example by A/B testing search headlines. The trap is the company that does neither, spending around $100,000 a month inefficiently without running experiments. He judges readiness by looking at funnel conversion, retention curves, LTV and churn, plus runway and monthly burn.

For prosumer premium SaaS in the mold of Grammarly, Canva or Whimsical, Timen says a healthy website-visit-to-free-account conversion is around 20% to 35%. Early on, with a small segment and strong product-market fit, it can reach 40% to 50%, but it tends to level off at 25% to 30% at scale. On free-to-paid, anything under 5% will not work long term, and he wants more than 5%, ideally above 7%.

Onboarding is almost always a big opportunity in this space, he says, because products like Airtable, Canva and InVideo are robust tools that democratize skills once held by professionals, and it is easy to get lost in their editors. Collecting relevant information up front lets a team tailor the landing experience and avoid exposing users to everything at once. Early-stage companies can often get two to four times better activation, while later-stage ones past Series B may still see a 20% to 30% lift.

Timen says getting to know your customer through interviews, surveys and panels always pays off, especially from seed to Series B, because it galvanizes a team with focus and clarity. Attribution investment is the opposite: companies tend either to underinvest and drift on inertia, or to chase scientifically pure measurement whose payoff may not materialize. The right level depends on stage.

Timen says TikTok is increasingly working well and sometimes becomes the most efficient digital channel for a brand. He pushes back on the idea that it is only for 15 to 22 year olds: the platform is big and still relatively unsaturated with advertisers. He has worked with brands whose core demographic is 40-plus, married and earning $200,000 or more in household income, and they found their audience there.

He also thinks offline channels such as podcasts and direct mail deserve more attention. They were long written off because online attribution was far more reliable, but offline measurement has improved while online attribution has deteriorated, so that gap has narrowed. He says podcasts are performing well for many brands, and Lenny adds that an investment of his saw a 10-to-1 return on a direct mail campaign.