Hello 👋,
Welcome to this week’s edition of Open Scout.
When you’re building in a crowded market, the natural move is to try to be better than the incumbents. Better features, better UX. This doesn’t work, and I think the reason is worth understanding beyond the obvious “they have more resources” explanation. The deeper problem is that “better” accepts the incumbent’s frame. You’re playing their game, by their rules, on a field they built. Even if your product genuinely is better on some axis, the customer has to notice, care, evaluate, and switch, and inertia is enormous. Most people don’t switch to something marginally better. They switch to something that solves a problem they can’t solve any other way.
Before we begin... a big thank you to this week's sponsor.
Open Scout Exclusive: You Could Be an Investor in High-Potential AI Infrastructure Startups Like These
Open Scout and Alumni Ventures are teaming up this week only to give readers early access to high-potential AI Infrastructure startup opportunities, including some of today’s most exciting AI Infrastructure companies co-invested alongside top VC firms like NVIDIA, BlackRock, and Andreessen Horowitz (a16z).
You get:
Curated deal flow of high-potential AI Infrastructure startups
AV is already investing alongside elite lead venture firms in these deals
No cost to see deals
No obligation to invest
Don’t miss your chance before access closes.
Join Alumni Ventures AI Infrastructure & Tools Syndicate Today →
If you’re interested in sponsoring this newsletter, please reach out here.
The companies that win in crowded markets tend to do something that looks, from the outside, almost timid. They pick a tiny slice of the market and try to own it completely. Amazon sold books for years before it sold anything else. Not because Bezos was passionate about literature. Books had predictable logistics, near-infinite selection, and universal demand, which made them the perfect proving ground for the fulfillment and trust infrastructure that Amazon would eventually use for everything. The books were the beachhead.
Klaviyo did this in email marketing. Mailchimp owned email marketing. Massive brand, hundreds of millions in revenue, entrenched everywhere. Klaviyo didn’t try to build a better Mailchimp. They built email specifically for e-commerce merchants, deeply integrated with Shopify, optimized for the exact workflows that online stores need: abandoned cart sequences, post-purchase flows, product recommendation emails. They weren’t competing with Mailchimp. They were competing for a customer Mailchimp was underserving. Klaviyo is now worth over $5 billion. Mailchimp is still around. They just don’t own e-commerce email anymore.
Veeva did something similar in CRM. Salesforce was the incumbent, obviously. Veeva didn’t build a better general CRM. They built a pharma-specific CRM on top of Salesforce’s own platform, with compliance features, clinical trial tracking, and regulatory workflows that pharma companies needed and Salesforce would never prioritize. That specificity took them all the way to IPO and a $25 billion market cap. They won by being the only option for pharma, not the best option for everyone.
DoorDash went to suburbs while Uber Eats and Grubhub fought over Manhattan and San Francisco. In suburbs there was no competition, everyone had a car so driver supply was easy, order values were higher, and your only alternative was Domino’s. Facebook was Harvard-only for months before expanding to other Ivy League schools, then colleges generally, and didn’t open to the public until late 2006, over two years after launch. Uber started as a black car limo service only in San Francisco in 2010. Not rides for everyone. Luxury rides for tech workers in one city.
None of these companies launched as the generic version of what they eventually became. I don’t think they could have.
Who exactly are you selling to
I’ve noticed that most founders, when asked about their target customer, give an answer that’s too vague to be useful. “SMBs.” “Mid-market SaaS companies.” “Healthcare.” These aren’t ICPs. They’re census categories. They describe millions of companies with nothing in common except a rough size or industry label.
A real ICP looks more like: “Series A SaaS companies with 20-50 employees who just hired their first sales team and are drowning in manual lead qualification because their CRM is a mess and the CEO is still doing half the selling.” That’s a group with a shared workflow, a shared pain, and a shared trigger event (the first sales hire) that makes them ready to buy right now, not someday.
The components that matter, I think, are these:
A trigger event, not just demographics. A construction firm that just won a government contract has different software needs than one that’s been doing residential for a decade. A SaaS company that just hired its first compliance officer is suddenly buying security tools it ignored for three years. The trigger separates “could use this someday” from “needs this now,” and the difference in conversion rate between those two states is enormous.
A workflow that existing tools handle badly. The best niches are defined by friction, not by industry. If every company in the niche hits the same wall at the same point in their process, that wall is what you build against.
People who actually talk to each other. This one is underrated. Veterinarians know other veterinarians. They go to the same conferences, read the same publications, complain about the same software. If your product is good, it will spread through a community like that almost by itself. In a scattered horizontal market, your users don’t know each other. Every customer is a cold acquisition. That’s expensive and it doesn’t compound.
A price anchor tied to pain, not to other software. This is maybe the most important one and I rarely see founders think about it explicitly. If your product replaces work that costs $150 an hour, you can charge $500 a month and the customer feels like they’re getting away with something. If you’re replacing a $20/month tool, you’re fighting over scraps. The best markets are the ones where you’re competing against expensive human labor, not against cheap software.
When you get the ICP right at this level of specificity, a few things start working that don’t work in a broad market. Your CAC drops because you can find your customers through their own channels. Your product gets better faster because every conversation teaches you something a generalist competitor will never learn. And word of mouth actually functions, because tight professional communities are where recommendations travel fastest. A physical therapist who saves 10 hours a week on documentation tells every PT she knows. Nobody tells their friends about a generic productivity app.
There’s a related thing I’ve been thinking about that might matter even more than the ICP itself: who’s building the company. The best wedge strategies aren’t just market strategy. They’re founder-market fit. A former clinic administrator building practice management software has advantages that are almost unfair. She already knows the workflow. She has the network. She has instant credibility when she walks into a practice and says “I ran one of these for eight years before building this.” She doesn’t have to spend six months learning what the office manager actually does, because she was one. The generalist founder starts from zero on all of that. They can learn it, and some do, but the gap is real and it shows up in everything from the product to the sales cycle to the referral rate. I think founder-market fit might be the most underweighted factor in whether a wedge strategy actually works. The wedge itself can be right and still fail if the founder doesn’t have the credibility or the knowledge to own it.
The framing game
I think the biggest positioning mistake founders make in crowded markets, especially in AI right now, is competing on features. “We do what they do, but we also have this.” You’ve already lost if that’s your pitch, because you’ve let the incumbent set the terms of comparison. The customer evaluates you feature-by-feature against the thing they already know, and the thing they already know wins on familiarity and switching costs even if it’s worse.
Positioning isn’t really about what your product does. It’s about what box the customer puts you in when they’re deciding. Change the box and you change everything.
I think the biggest positioning mistake founders make in crowded markets, especially in AI right now, is competing on features. “We do what they do, but we also have this.” You’ve already lost if that’s your pitch, because you’ve let the incumbent set the terms of comparison. The customer evaluates you feature-by-feature against the thing they already know, and the thing they already know wins on familiarity and switching costs even if it’s worse.
Positioning isn’t really about what your product does. It’s about what box the customer puts you in when they’re deciding. Change the box and you change everything.
Consider a database product struggling against Oracle. Feature-by-feature against Oracle is suicide. But the same product repositioned as a business intelligence tool suddenly had a completely different competitive set. The speed that was irrelevant in the database category became the key differentiator in the BI category. Nothing about the product changed. The context changed, and that was enough.
In AI this is especially important because there’s a trap that almost everyone falls into. You say “we use AI to do X.” That positions you in the AI category. The customer’s mental model becomes “AI tool,” and in that frame, ChatGPT and Claude win every time because they’re the known quantities with the biggest brands. The repositioning move is to stop being an AI tool and start being the outcome your customer actually wants. Not “AI for legal research.” “The system that cuts deposition prep from 40 hours to 4.” Not “AI-powered sales development.” “The tool that books qualified meetings while your team is asleep.” Different frame, different competitive set, and crucially, different price anchor. You’re no longer competing against $20/month software. You’re competing against a paralegal’s billable rate or an SDR’s fully-loaded salary.
I suspect a lot of AI startups that are struggling right now could fix their growth by changing nothing about the product and changing everything about how they position it.
Growing from the wedge
The wedge1 gets you in. But you can’t stay in the wedge forever, or you end up as a nice small business in a tiny niche, which is fine if that’s what you want but probably isn’t why you’re reading this.
Two expansion paths. Vertical first, horizontal second. The order matters and most founders get it wrong.
Vertical expansion means selling more products to the same customer. You started with bid management software for commercial contractors. Now you add project scheduling. Then material procurement. Then subcontractor payment tracking. Then safety compliance. Each new product costs almost nothing to acquire because you already own the relationship. And here’s the part that I think is underappreciated: each additional product makes the customer dramatically harder to lose. Replacing one tool is a Tuesday afternoon project. Replacing five integrated tools is a quarter-long migration that nobody will volunteer for.
Veeva did this. Started as a pharma CRM. Then added Vault, a content management system for regulatory submissions. Then clinical data management. Then safety reporting. Each product served the same pharma customer, deepened the relationship, and made switching away from Veeva an enterprise-wide disruption rather than a department-level decision. Gorgias did the same in e-commerce support: started as a helpdesk for Shopify merchants, then added revenue generation, automation, loyalty tools. The merchant using Gorgias for support, revenue, and loyalty is locked in by the integration, not by any single feature.
Horizontal expansion means taking what works in your niche and bringing it to adjacent ones. You built scheduling software for physical therapy clinics. Chiropractic offices have similar appointment flows, similar insurance billing, similar patient communication needs. Occupational therapy too. The test is whether the product, the go-to-market, and the domain knowledge transfer without rebuilding from scratch. If they do, expand. If you’d basically be starting over, don’t, because you’d be abandoning the wedge’s advantages and competing generically in a new market where you have no edge.
Do vertical until the position is strong enough that nobody can take your niche while you’re looking the other way. Then go horizontal. Not before.
I should be honest about the risk here, because the pro-niche argument has a failure mode that doesn’t get discussed enough. Some companies niche too hard and can’t get out. AbeBooks built a great business in used and rare books. When they tried to expand into general books, Amazon was already there and impossible to dislodge. They ended up selling to Amazon in 2008 for a fraction of what a more expansive company could have been worth. The niche that was supposed to be the beachhead became the ceiling. This happens when you wait too long to expand, when the adjacent market gets taken while you’re perfecting your niche, or when your niche is so specialized that nothing you’ve built transfers to the next market.
The timing question is genuinely hard. Expand too early and you dilute the wedge before it’s won. Expand too late and the adjacent territories get claimed. The signals I’d watch for: when your niche penetration is high enough that growth is slowing not because you’re doing something wrong but because you’re running out of prospects. When customers start asking for things that naturally extend into adjacent territory. When a competitor enters your niche and you can defend it without all-hands effort, meaning the position is strong enough to hold while you look outward.
When the incumbent notices you
At some point in every successful wedge strategy, the incumbent wakes up. Salesforce noticed Veeva eating pharma. Mailchimp noticed Klaviyo eating e-commerce. Zendesk noticed Gorgias eating Shopify support. The question is what happens next.
Usually, the incumbent does one of three things: they build a competing vertical product, they try to acquire you, or they dismiss the niche as too small to matter. The third option is the best one for you, because by the time they realize the niche was actually the beachhead for a much larger business, your domain depth is years ahead of anything they could build. Veeva had spent years building pharma compliance features and earning FDA-adjacent credibility that Salesforce couldn’t replicate with a six-month product sprint. Klaviyo had years of e-commerce behavioral data and Shopify integration depth that Mailchimp couldn’t match by adding a Shopify plugin.
The first option, building a competing vertical product, is the most common response and usually the least effective. Incumbents are bad at verticals because their entire organization is built for horizontal scale. Their sales team sells to everyone. Their product team prioritizes features that serve the broadest base. Dedicating serious resources to one vertical feels like a step backward to a company that’s been growing by going wide. They’ll launch something, it’ll be mediocre compared to yours, and their own sales team won’t prioritize it because the commissions are smaller than the horizontal deals.
The real danger is the second option: acquisition. And honestly, if an incumbent offers to buy you, that’s not a failure. That’s validation that the wedge worked. Your job is to make sure you’re far enough along that the acquisition price reflects the business you’re building, not just the niche you’ve won.
How to actually get customers
Most founders conflate demand capture and demand creation, which is a mistake in any market but an expensive one in a crowded market.
Demand capture is converting people already looking. Google searches, review sites, comparison shopping. High intent, converts well, but in a crowded market every competitor is bidding on the same keywords and the auctions get expensive fast. You can win here but it’s a spending war and the incumbents usually have deeper pockets.
Demand creation is reaching people before they start looking. Content, community, founder-led outreach, events. Lower intent but much cheaper and, in a niche, incredibly efficient. A piece of content about “how mid-size accounting firms are automating 60% of their audit prep” reaches every managing partner thinking about margins during busy season. Specific content in a concentrated community is the highest-leverage marketing a startup can do. Generic content in a broad market reaches nobody because it’s relevant to nobody in particular.
And then there’s a third thing that I think is more powerful than either and weirdly underused: distribution leverage. Partnering with someone who already has the relationship you need. An industry association. A PE firm rolling up businesses in your vertical. A consultancy that already serves your ICP. A complementary software tool. If you’re selling to independent real estate brokerages and you partner with the National Association of Realtors, you’ve accessed 1.5 million members without buying a single ad. That’s distribution no competitor can replicate with spend.
Why knowledge beats technology
I want to end on something that I think is the most important idea in the essay, even though it sounds obvious when you say it directly. In a crowded market, the hardest thing to replicate is not your technology. It’s what you know.
Every competitor has access to GPT-5 or Claude or Gemini or whatever the frontier model is next month. The foundation models are a commodity. The cloud infrastructure is a commodity. Most of the open-source tooling is a commodity. What isn’t a commodity is the understanding you build by spending months inside your customers’ actual workflows.
Six months inside freight brokerage operations teaches you that brokers spend half their day re-keying data between TMS, carrier portals, and customer spreadsheets. That the mid-size brokerages running 500-2,000 loads a month have completely different tech stacks than the enterprise shops. That the ops manager, not the CEO, decides which tools get adopted. That FMCSA compliance documentation is a constant low-grade headache nobody talks about because it’s just how things are. A competitor spread across ten industries will never learn any of this because they’re not paying close enough attention to any single one.
And that knowledge shows up in the product in ways the customer can feel. Features designed by someone who understands the workflow from the inside feel different from features designed by someone who read a TAM slide in a pitch deck. The customer can tell. And their recommendation, the one that actually drives growth, is “this was built by someone who gets it.” That’s not something you can manufacture with marketing.
The moat gets deeper every month you spend in the niche. That’s the opposite of technology advantages, which erode as competitors catch up. Domain knowledge compounds. It’s the wedge within the wedge. And in a market where every competitor has the same AI, the same cloud, the same tools, it might be the only durable advantage left.
Thanks for reading.
If you enjoyed this issue, send it to a friend—it helps more than you think.
Back in your inbox Wednesday,
A wedge is simply a strategy to win a large market by initially capturing (1) a tiny part of a larger market or (2) a large part of a small adjacent market.