1. 🧭 Getting Noticed in the Most Crowded Field in This Series
Part 6 built a product; Part 7 covered the research culture behind it, if any. This article is about the much harder, less controllable problem of getting anyone to notice, especially in Lane 2, which our earlier Startups-to-Watch analysis already identified as the most crowded lane in this entire market. "AI startup marketing" advice from even a year or two ago is often stale — differentiation now has to happen at the positioning and distribution level, not just the "we use AI" claim itself, since virtually every competitor makes some version of that claim too.
10M → 30MPerplexity's monthly user growth from January 2024 to April 2025 — 3x in roughly fifteen months
400%Perplexity's year-over-year revenue growth, $20M to $100M, in the same broad window
159MPerplexity's reported monthly users by 2026 — achieved without a traditional ad-led playbook as the primary driver
Multi-ChannelPerplexity's actual mix — OOH, creator partnerships, social, in-app prompts, and enterprise partnerships together, not one silver-bullet channel
2. 📣 Go-to-Market, By Lane
🔴 Lane 1: Frontier Scale-First
"Marketing" here is closer to research and safety credibility signaling than a traditional consumer or B2B campaign — a benchmark result, a safety research publication, or a high-profile researcher joining the team functions as the primary marketing event, aimed at investors, potential hires, and enterprise partners rather than end consumers directly. Product marketing in the traditional sense mostly arrives later, once an actual product exists to market — the earliest GTM motion is really about maintaining the research credibility and hiring pipeline covered in Part 7.
🔵 Lane 2: Applied / Agentic Layer
This is where a genuine go-to-market strategy matters most and is hardest to get right, precisely because the lane is so crowded. Product-led growth (letting the product itself, used well, drive word-of-mouth and adoption) has become the dominant successful pattern, but it requires the product to actually be good enough to generate organic sharing — a GTM strategy can't compensate for a product that isn't there yet. Community and developer-channel presence (technical content, open-source components, visible presence where your specific customer type already spends time) tends to outperform paid acquisition at this stage, both on cost and on credibility with a technically sophisticated buyer.
⚪ Lane 3: Narrow Research Bet
There is, functionally, no product marketing — the entire "marketing" motion is publishing research, maintaining visibility in the specific technical community that cares about your thesis, and managing investor and future-hire perception of the company's credibility. This is the inverse problem from Lane 2: rather than differentiating a product in a crowded market, the challenge is making a genuinely novel, hard-to-explain technical bet legible enough that the right people (specialized investors, top researchers who might join) understand why it matters before any product exists to demonstrate it.
3. 📡 Real Channels That Actually Work in 2026
🔄 Product-Led Viral Loops
In-app sharing prompts, referral mechanics, and a product experience good enough that users organically recommend it — Perplexity's core growth engine, and the most defensible channel because competitors can't simply copy a genuinely good product experience overnight.
🤝 Strategic Distribution Partnerships
Partnering with a company that already has the distribution you need — Perplexity's SoftBank partnership for Japanese enterprise sales, and device-integration deals with Samsung and Motorola, are examples of borrowing someone else's existing sales motion or install base rather than building distribution from scratch.
🔌 API-as-a-Service as a Growth Channel
Perplexity's Sonar API lets other developers embed its core capability into their own products — a distribution channel that scales through other people's products, not just your own, and generates usage and revenue simultaneously.
4. 🎯 Positioning — Saying Something Competitors Can't Say
Generic "AI-powered" positioning has been fully commoditized by 2026 — the same messaging every Lane 2 competitor uses, since virtually all of them are built on the same handful of frontier model APIs (per this series' own escalation-ladder framework in Part 6). Effective positioning in a crowded lane comes from specificity: naming the exact customer, the exact workflow, and the exact outcome your product delivers, rather than a broad capability claim. This connects directly back to Part 1's validated wedge — a company that skipped narrow validation in favor of broad ambition almost always ends up with generic, forgettable positioning as a direct consequence, not a separate marketing failure.
5. 🏛️ Case Study: Perplexity's Multi-Channel Growth Engine
Perplexity AI
10M → 159M Monthly Users
3x growth in ~15 months400% YoY revenue growthSoftBank, Samsung, Motorola partnerships
Perplexity grew from 10 million monthly users in January 2024 to 30 million by April 2025 — a 3x increase in roughly fifteen months — alongside 400% year-over-year revenue growth ($20M to $100M), reaching a reported 159 million monthly users by 2026. Rather than relying on a single growth lever, the company ran a genuinely multi-channel strategy simultaneously: out-of-home advertising, creator partnerships, social content, and in-app referral prompts working together across both discovery and retention; a strategic partnership with SoftBank specifically to access its existing enterprise sales team in Japan rather than building a Japanese sales org from scratch; device-level integration deals with Samsung and Motorola putting the product directly in front of users who never searched for it; and a Sonar API-as-a-service channel letting other developers embed Perplexity's core search capability into their own products, generating distribution and revenue simultaneously.
The lesson for this article: there was no single "growth hack" — the actual strategy was running several structurally different channels (viral product loops, borrowed distribution via partnerships, and a developer-facing API channel) at once, so that the company wasn't dependent on any one channel continuing to work. A Lane 2 founder looking for one silver-bullet channel is asking the wrong question; the right question is which combination of channels fits the specific product and customer type.
6. 📋 Side-by-Side: Go-to-Market by Lane
| Factor | 🔴 Lane 1: Scale-First | 🔵 Lane 2: Applied Layer | ⚪ Lane 3: Research Bet |
|---|
| Primary "marketing" activity | Research/safety publications, benchmark results | Product-led growth, content, partnerships | Research publication, technical community visibility |
| Target audience | Investors, top research talent, enterprise partners | End customers or developers directly | Investors, specialized research peers, future hires |
| Effective channel example | A widely-discussed benchmark or safety paper | Viral product loops, developer communities, strategic distribution partnerships | Peer-reviewed or widely-cited research output |
| When traditional marketing starts | Only once a real product exists | From day one — it's core to the business | Possibly never, if the thesis never becomes a product |
| Biggest positioning risk | Overclaiming capability ahead of what's actually shipped | Generic "AI-powered" messaging indistinguishable from competitors | Thesis too technical/niche for non-specialist investors to grasp |
7. ⚠️ Risk Flags
🌫️
Generic "AI-Powered" Positioning
In a lane this crowded, positioning language that any of a hundred competitors could use verbatim is functionally invisible — specificity about customer, workflow, and outcome is the only real differentiator left at the messaging level.
📢
Marketing Ahead of Product Readiness
Especially risky in Lane 1, where a benchmark claim or capability announcement that outpaces what's actually reliably shipped creates a credibility gap that's expensive to repair — the ARC-AGI harness controversy covered elsewhere in this site's research is a real-world instance of exactly this risk playing out publicly.
🎯
Chasing One Growth Channel Exclusively
Perplexity's multi-channel approach is instructive precisely because a single-channel dependency (all paid ads, or all one partnership) creates fragility if that channel's economics or availability changes.
🔬
Confusing Research Visibility With Product Marketing
Specific to Lane 3 — publishing genuinely interesting research doesn't automatically translate into the kind of business-facing credibility needed for a future product launch; the two audiences and messaging approaches are different and require deliberate translation when the time comes.
Perplexity didn't win with one growth hack — it ran viral product loops, borrowed distribution, and an API channel simultaneously, so no single channel's failure could sink the whole strategy.
8. 🧪 Go-to-Market Checklist (All Three Lanes)
1
Confirm your product is actually good enough to generate organic word-of-mouth before investing heavily in growth channels — no GTM strategy substitutes for a product people genuinely want to share.
2
Write positioning specific enough that a direct competitor couldn't use the same sentence verbatim — if they could, it's not differentiated yet.
3
Identify at least two structurally different channels (not two variations of the same channel) to reduce single-point-of-failure risk, following Perplexity's multi-channel pattern.
4
For Lane 1/3, treat every research publication or benchmark claim as a marketing event — plan its framing and audience deliberately, rather than treating publication and marketing as separate, disconnected activities.
5
Match claims to actual, verifiable capability — the credibility cost of an overclaim that gets publicly scrutinized is far higher than the short-term attention it might generate.
9. 🧭 What's Next in the Series
Part 9 covers Sales & Early Customers — from first deal to repeatable revenue, again across all three lanes: what "closing a deal" even means ranges from a Lane 2 self-serve signup to a Lane 1/3 enterprise or strategic partnership negotiated over many months.