The a16z Show Analysis: Lighthouse vs. Landgrab—How to Choose an AI Sales Strategy (Guests: Joe Schmidt, Andy)
August 13, 2026
Overview
In this episode of the a16z podcast 'The a16z Show,' Joe Schmidt of a16z discusses the two sales strategies that enterprise AI startups should adopt—'Lighthouse' and 'Landgrab'—based on his article 'Lighthouse or Land Grab.' He is joined by guest Andy, who has built sales organizations at Samsara and Meraki, to discuss how to choose a strategy while sharing insights from their own launch experiences.
Joe Schmidt says that while driving on Highway 101 in San Francisco, he noticed that competing startups selling the same software on both sides of the road were all targeting the same famous San Francisco companies. From bus advertisements to airplane banners, everything was aimed at the same customer segment, leading to a hyper-concentration of sales efforts.
He points out that it is not always necessary to chase famous companies in San Francisco or New York; there is also the option of finding customers who need the product in places like Ohio, Chicago, or St. Louis. The two playbooks, Lighthouse and Landgrab, organize these two strategies: 'chasing prominent logos (customer names)' versus 'quickly capturing the entire market.'
Two Strategies and a 2x2 Matrix
Joe Schmidt suggests organizing deals using a consulting-style 2x2 matrix.
The vertical axis is 'Buyer Exposure.' This refers to the risk if the purchased solution fails, the degree to which the product is exposed to the company's end customers, and the overall risk associated with the software purchase. The horizontal axis is 'Whether the proof propagates within the market.'
The top-right quadrant is 'High exposure, high proof propagation market,' which is the Lighthouse market. Regulated industries are typical here; the number of customer logos is limited, and buying the wrong software could lead to scrutiny from regulators.
The bottom-left quadrant is 'Low exposure, low proof propagation market,' which is the Landgrab market. This is an area where budgets are already established and buyers have been paying for specific services, allowing new entrants to show the math that they are better than the current solution (whether software or human-based). In short, the top-right prioritizes track records and social proof, while the bottom-left prioritizes numbers (effectiveness).
Examples of Lighthouse and Landgrab
In the show, Hebbia and Harvey were cited as typical examples of the Lighthouse strategy, while Stu and Decagon were cited as typical examples of the Landgrab strategy, as discussed in the article.
Harvey is an AI for the legal field that automates tasks routinely performed by junior lawyers, targeting an area that was extremely new and high-risk at the time. By acquiring a few key law firms early on, their track record propagated significantly within the market, spreading the perception that 'it is safe to buy this solution.'
Meanwhile, Stu is a Landgrab-type company that is rebuilding accounts receivable (AR) management with AI. Founders Tark and Ben replaced the debt collection work previously done manually by human teams and software with a model where AI and humans collaborate. By showing the math that they have higher collection efficiency than current solutions or human collection teams, they were able to conduct sales that forced mid-market customers to give an immediate answer. The proposal improves working capital significantly, contributing to both cost reduction and increased profits.
Andy also cited AI-native customer support companies Decagon and Pylon as good examples of Landgrab. Decagon clearly claims to 'achieve better customer support than others' and gains trust by achieving pre-agreed benchmarks within a set period. On the other hand, Further AI (where Andy serves as a board member), which brings AI into the insurance field, was introduced as an example of a Lighthouse strategy that puts governance and security at the forefront when dealing with some of the world's largest insurance companies.
Samsara, Catching the Tailwinds of the ELD Mandate
Andy joined Samsara ($IOT) in 2017. Founded in 2015, Samsara's initial vision was to spread internet-connected sensors across every supply chain and provide data to companies. The first to become popular were telematics terminals.
In long-haul trucking, paper logbooks were used before 2016. The ELD mandate (Electronic Logging Device mandate) began to be introduced in the U.S. around 2016 and was implemented in phases from 2016 to 2019. This created a situation where almost all trucking companies suddenly needed to secure a budget for this area, creating demand to 'try new options.'
Although major companies with hundreds of millions of dollars in annual revenue, such as AT&T ($T) and Verizon ($VZ), had already entered the market, the entire market became a tailwind for the emerging Samsara. According to Andy, in the early days of the company, rather than strategically choosing between Lighthouse or Landgrab, they moved forward by asking customers, 'Who will pay for this?'
Meraki, Landgrab with Free Access Points
Andy continued by reflecting on Meraki's journey as a cloud networking company. Meraki was founded in 2006 by PhD students from MIT and was originally a research project for roof-mounted mesh Wi-Fi called Roofnet. Initially, it was a Wi-Fi business for municipalities and parks, but after a few years, the difficulty of the business model became apparent, and they pivoted to the enterprise sector.
In the enterprise networking market around 2009-2010, Cisco ($CSCO) and HP dominated the major players, and it was considered impossible for a startup to break in. So, Meraki adopted a landgrab strategy. They sold the value of being "easy to set up, fast to deploy, and simple to manage" not to large enterprises, but to IT teams in the small-to-mid-market.
To achieve this, they implemented a policy of sending a free access point to anyone who attended a webinar. By having customers try it out and experience the simplicity compared to existing Cisco products, they were able to achieve long-term results. Andy recalls, "The fundamental approach was to have customers experience the technology and realize it was superior to the alternatives."
Pitfalls of ACV and Trial Design
Regarding the sales team's ACV (Annual Contract Value), Andy stated, "As long as you don't break your unit economics, if the conditions are met, don't worry about the small numbers; just close as many deals as possible." The idea is that if your system is built to run on $15,000 ACV deals (approx. 2.39 million yen at 159.29 yen/dollar), it is better to acquire a large volume of $15,000 deals rather than chasing $8,000 deals (approx. 1.27 million yen at 159.29 yen/dollar). He suggests that gradually moving up to larger enterprises is the natural progression.
In the AI field, trials and proofs of concept are becoming prolonged, increasing the risk of them becoming "science projects." Andy emphasized the importance of strictly defining durations—such as 30, 45, or 60 days—and defining success criteria in advance. While the duration varies depending on the complexity of the product, it is essential to clearly define both the end date and the success criteria.
Joe Schmidt pointed out that because AI is highly configurable, even if the product works correctly, it takes time to see results if the customer's usage and optimization are not aligned. He stated that the sales side needs to distinguish between "the product working" and "the customer using it correctly," and that it is necessary to educate the customer on what will be proven before signing the contract.
Vertical Expansion and Transition to Lighthouse
According to Andy, there are almost no examples of companies that have become large enterprises without using both strategies. Even if they start with a landgrab in the early stages, they transition to lighthouse as they mature, and conversely, they may acquire large enterprises through lighthouse and then expand to the entire market. Both Meraki and Samsara started with landgrab and then expanded vertically.
For Meraki, school districts were the first lighthouse market, while for Samsara, it was the public sector, such as municipalities, counties, and states. Because school districts talk to each other frequently, once you acquire the largest school district in each state, social proof spreads rapidly to the districts below them. Since the sales cycle, decision-makers, and procurement methods in the public sector differ from regular enterprise sales, it is necessary to have a dedicated sales team to handle them.
From PLG to the Phase of "Selling Big Software Again"
In the second half of the show, developer-led bottom-up adoption and PLG (Product-Led Growth) were also discussed. Andy stated that while PLG still exists and there are success stories like Cursor, the landscape has changed.
For the 10 to 15 years prior to 2024, large cloud infrastructure companies like CRM, HR, ITSM, and security had already solidified the market, so the only way to get into an enterprise was to start with a small wedge product and expand. However, simple cloud-to-cloud migrations were unlikely to happen, leaving PLG as the only viable path.
Currently, AI is fundamentally changing core software like CRM and HR. Joe Schmidt said that in the trend where "humans do higher-value work and agents handle routine tasks," the moment to sell large-scale software has returned. Buyers are more educated than ever in the last 15 years and clearly understand what they want. The job of the sales side has shifted from long-term education to "convincing them that your company is the answer."
Common Founder Mistakes and Practical Advice
Joe Schmidt stated that the biggest mistake founders make early on is overthinking strategy. Instead of spending time on strategy, he suggests spending 1% on strategy and 99% on execution, talking to customers, and chasing "people who will buy the product today." Just because you land a famous company doesn't mean there is a special multiplier on revenue; it is important to chase the customers you can acquire and keep improving the product.
Andy agreed with this and advised against falling into analysis paralysis.
Lightning Round: Memories of Deal Locations, Sales Organization Wisdom
Finally, Joe Schmidt asked Andy a few short questions.
The most unusual places I've closed deals include fishing trips, shooting ranges, ballparks, Applebee's, truck yards, and sanitation facilities. The deals at Applebee's have totaled millions of dollars (several hundred million yen at 159.29 yen/dollar).
My advice to my younger self is: 'Don't look at salary, commission, or job title; look for the company where you will grow the most.'
The role you should hire for in a sales organization earlier than usual is the first Sales Operations or Revenue Operations person. It is crucial to have someone who can organize sales representative deployment, rosters, commission structures, and sales policies.
Andy asserts that an early-stage sales team should aim for a 100% quota attainment rate. If only 40-50% of the team is hitting their quota, it means either the quotas are too high or the hiring profile is off. He argues that in the early stages, you should prioritize building momentum over sales costs.
The discussion between the two a16z guests demonstrates that even in the AI era, the principles of 'which market to enter and how to sell' remain continuous with past experience in the software industry.
▶ Previous entry in this series: 'Integrity' and 'Courage' Required for Founders in the AI Era—Garry Tan Discusses New Rules https://note.com/yondo/n/n41b01bc4c8eb
▶ Read more articles on this theme: Magazine 'The Current State of AI and Technology' https://note.com/yondo/m/m23cc16367f2e
#AIDevelopment #Podcast #a16z #Samsara #IOT #Cisco #CSCO #AT&T #T #Verizon #VZ #HP #JPMorganChase #JPM #SalesStrategy #GenerativeAI #EnterpriseAI #Startup #Lighthouse #Landgrab #PLG
