Selling Before Building: $1M ARR in Six Months

概览

This episode features Omar Khan interviewing Julius Kurfgen, co-founder of Uplane, an AI-driven marketing platform focused on automating research, ad creation, landing pages, campaign operations, and performance iteration.

The core theme is “sell first.” Julius argues that founders should begin with discovery calls and paid customer commitments before writing code, then rapidly build working demos around the problems customers describe.

The discussion also covers Uplane’s hybrid model: a software company that often delivers outcomes through a managed-service layer, especially while automation is still being built out. Julius explains why Uplane uses performance-linked pricing, how it handles attribution, and why enterprise customers need guardrails before allowing AI into brand-sensitive workflows.

分段落总结

[00:00] Opening Thesis: Sell Before Building

[事实] Julius says founders should start selling before touching code: run discovery calls, learn customer problems, schedule a demo one week later, and build something working in that sprint. [事实] The host introduces Uplane as a company that reached its first million in ARR about six months after launch. [事实] The episode promises discussion of LinkedIn cold outreach, one-week demo sprints, avoiding free pilots, and replacing seat-based pricing.

[02:07] What Uplane Does

[事实] Uplane targets wasted spend in digital marketing, where Julius says only around 10% of ads are actually profitable. [事实] Julius says companies keep running underperforming ads because of slow manual processes, design bottlenecks, and coordination across agencies or in-house teams. [事实] Uplane automates marketing research, ad and landing-page production, cross-channel publishing, monitoring, and doubling down on winning assets while taking losers offline. [推测] Uplane’s core positioning is less about creating more ads and more about faster experimentation tied to performance data.

[03:31] Ideal Customer and Company Size

[事实] Uplane works with companies using major platforms including Google, Meta, TikTok, LinkedIn, Taboola, Opera, and others. [事实] Julius says the product works best for companies spending more than $100,000 per month on ads because there is enough budget to test and iterate quickly. [事实] Uplane is operating at a seven-figure ARR level, has about 20 people, and has teams in San Francisco and Berlin. [事实] Julius names Kalshi and AG1 as U.S. clients.

[05:03] Origin of the Product

[事实] Julius’s background is in marketing, and he previously worked as marketing director at NPAW, a European startup or scale-up. [事实] He says he personally built around 10,000 ads by hand and experienced the pain Uplane now solves. [事实] His team struggled to track ads cleanly, analyze what headlines, CTA buttons, and image components worked, and iterate fast enough. [推测] The product idea came from founder-market fit: Julius had lived inside the workflow rather than only observing it from outside.

[07:19] Why the Founders Committed

[事实] Julius emphasizes that Uplane was built with his two co-founders, not by him alone. [事实] He says all three founders had long wanted to build a company. [事实] He viewed marketing operations as a large business problem because marketers want to tell stories and execute ideas but often get stuck in repetitive work and messy tools. [推测] The founding motivation combined personal frustration, a large market pain, and a belief that AI could remove repetitive marketing work.

[08:09] Validation Through Discovery Calls

[事实] Julius had strong conviction from the start because he would have bought the product in his previous role. [事实] His former team later became a Uplane client, but not the first client. [事实] Uplane’s first customers came through LinkedIn outreach and discovery calls with smaller clients. [事实] Julius says the purpose of these calls was to understand whether prospects worked with agencies, had internal design teams, struggled with creative production, or had pain in media buying.

[09:16] LinkedIn Outreach Method

[事实] Julius describes the early outreach as humble and direct: he had left his job, had an idea in the marketing space, and wanted to ask questions. [事实] He says people open up when founders approach calls with a learning mindset instead of trying to teach or sell immediately. [事实] After identifying a problem, Uplane would schedule a follow-up demo one week later and build a solution for that specific customer problem. [事实] Julius says their first customer was a Berlin startup reached through cold outreach, though he does not name the company.

[11:40] Paid Pilots and the “Sell First” Rule

[事实] Julius says his biggest advice is to start selling first, before writing code. [事实] He defines selling as running discovery calls, learning about problems, booking a demo a week later, and building something working in time for that demo. [事实] He argues that founders should avoid free pilots because a dollar sign is needed to validate whether the delivered value has a real business case. [推测] In Julius’s view, willingness to pay is a stronger validation signal than polite interest or unpaid usage.

[16:05] What Early Demos Looked Like

[事实] Uplane initially sold a service-like marketing outcome, aiming to do what agencies do but faster and with higher quality through AI. [事实] Early demos showed outputs such as a full marketing strategy, a new website, ad strategy, and persona-specific approaches. [事实] Julius says he would not recommend fake demos because AI now allows founders to build scrappy but real prototypes quickly. [事实] The host and Julius agree that founders should be clear when showing a prototype rather than pretending it is a fully working product.

[18:24] Founder Responsiveness as an Advantage

[事实] Julius describes using a very early-stage founder’s product despite bugs because the founder responded quickly and solved problems immediately. [事实] Uplane has a rule that client questions should be answered within 120 seconds. [事实] Julius says this level of reachability is an early-stage founder’s advantage over agencies that may take days to respond. [推测] Uplane treats speed of communication as part of the product experience, not just customer support.

[20:14] Software Company With a Managed-Service Layer

[事实] Julius says Uplane is a software company with engineers building marketing automation software. [事实] He says many clients do not want to use software directly; they want someone to run their marketing. [事实] Uplane therefore uses account managers and AI marketing strategists who operate the platform for clients, especially where automation is incomplete. [事实] The company also sells software directly to enterprise firms with large marketing teams that want to keep work in-house. [推测] The managed-service model helps Uplane sell before every part of the software is fully automated.

[23:20] Landing Deutsche Bahn

[事实] The host says Uplane landed Deutsche Bahn about nine months after launch. [事实] Julius says the relationship began through someone he already knew, using the same discovery-oriented approach as cold outreach. [事实] The first pain point discussed was an ad-formatting process where assets had to be sent to an agency and returned after about two weeks. [事实] Uplane showed that this formatting step could be automated, which opened conversations with creative, media buying, and analytics teams. [事实] Julius says the enterprise process involved many stakeholders, compliance, purchasing, and about nine months of conversations.

[28:00] Performance-Based Pricing

[事实] Uplane does not rely on seat-based pricing in the discussed model. [事实] Julius says Uplane charges a fixed fee to cover operating costs, with little or thin margin, plus a variable success-based fee. [事实] He says the pricing pitch is easier because Uplane earns meaningfully when the client earns or succeeds. [事实] He frames the model as incentive alignment and says it reduces risk for clients. [推测] The model shifts more risk to Uplane but can make adoption easier for customers entering a relationship with a young company.

[30:14] Attribution, Spend, and Reporting

[事实] Julius says Uplane generally charges a percentage of ad spend. [事实] The logic is that if clients are happy with Uplane-run campaigns, they will shift more budget onto them. [事实] He says most clients already have tracking in place, and Uplane also helps improve tracking and implement its own tracking technology. [事实] Clients receive daily reporting showing where Uplane-run spend goes, which ads are running, and the cost per result. [推测] Uplane avoids some revenue-attribution disputes by tying fees to managed ad spend and transparent campaign performance rather than claiming total business revenue impact.

[32:11] AI Volume Versus AI Quality

[事实] Julius criticizes the use of AI to send more low-quality, one-way spam faster. [事实] He says AI-driven outreach often uses surface-level personalization that does not understand the recipient’s real problems. [事实] He argues that AI must break down silos and access relevant information, such as analytics data about which ads actually perform. [事实] In marketing, this means generating new ads as iterations of proven winners rather than simply increasing throughput. [推测] Julius sees the strategic value of AI as decision quality and relevance, not raw content production.

[35:26] Finding the 10% That Works

[事实] The host frames the challenge as selecting the 10% of content that drives most results. [事实] Julius says high-performing ads depend on personas, use cases, timing, and context. [事实] He gives examples such as Deutsche Bahn targeting grandparents with messages about visiting grandchildren, or Kalshi producing sports-related ads for specific fan groups. [事实] Uplane’s goal is to bring the most relevant ad about a topic people care about to the right audience. [推测] Uplane’s advantage depends on matching creative variations to narrow audience intent rather than broadcasting generic ads.

[37:00] Brand Guardrails and Atomic Content

[事实] Julius says Uplane controls AI output through “atomic content,” using pieces supplied by clients such as brand guardrails, compliance rules, reference ads, product descriptions, ERP access, and past templates. [事实] Uplane trains or configures AI to assemble ads from approved resources and operate within strong guardrails. [事实] Julius says Uplane uses code and evaluations to stop AI from going beyond those constraints. [事实] For managed-service clients, account managers review ads before they go live. [事实] Uplane sometimes builds brand guideline snippets for clients to improve AI output.

[39:00] Personal Experience Versus Customer Discovery

[事实] The host notes that much of Uplane’s early product direction came from Julius’s own experience, while still using discovery calls. [事实] Julius says many founders have to pivot often, and he remains open to Uplane changing direction if a better opportunity appears. [事实] His default answer to customer requests in discovery is often “yes,” followed by finding an engineer and trying to build a solution. [事实] He says sales and product work very closely together, discussing pipeline clients, their issues, conversion likelihood, and demo readiness. [推测] Uplane balances strong founder conviction with ongoing customer-led iteration.

[42:20] Differentiation in a Crowded Market

[事实] Julius says Uplane’s differentiator is that it is not a point solution for only ads, analytics, or media buying. [事实] Uplane integrates analytics, competitor information, viral TikToks, asset creation, and media buying into one workflow. [事实] He says this integrated approach is difficult to build but is why clients choose Uplane. [推测] Uplane’s competitive claim is that intelligence improves when the full marketing workflow shares data across functions.

[44:23] Lightning Round

[事实] Julius names “sell first” as one of the best pieces of business advice he has received. [事实] He recommends The 7 Habits of Highly Effective People because it teaches proactivity and taking ownership. [事实] He says one of the best business expenses was paying for visas to bring German employees to the U.S. after Y Combinator and early U.S. clients. [事实] His favorite productivity habit is planning; he says he became highly organized to manage his naturally chaotic tendencies. [事实] Outside work, he enjoys being outdoors, swimming in San Francisco, cycling, and appreciating small moments in nature.

播客点评/总结

[事实] The episode is valuable for SaaS founders because it gives concrete go-to-market tactics: discovery-led LinkedIn outreach, one-week demo sprints, paid validation, fast customer response, and pricing tied to customer success.

[事实] A strong part of the conversation is that Julius explains Uplane’s messy early model honestly: the company sells software, but uses a managed-service layer and human operators where automation is not yet complete.

[推测] The main limitation is that the episode does not deeply quantify Uplane’s customer results, margins, churn, or exact ARR, so listeners should treat the growth story as directionally useful rather than a complete operating blueprint.

[推测] This episode is best suited for B2B SaaS founders, AI application builders, and marketing-tech operators who are figuring out how to validate demand, sell before product maturity, and build automation around a painful existing workflow.