Startup Growth: Why Early Sales Can Mislead
Learn why early startup sales can mislead, how to test repeatable demand in your first 90 days, and when to build a sales team. Use this practical framework.
Don't Scale Startup Sales Until You Can Repeat Them
Early revenue feels like proof. A new product launches, customers pay, and suddenly the team starts talking about hiring salespeople and pushing growth. But a handful of early deals can be misleading: those customers may have bought for very different reasons, found you through unrelated channels, or needed enough custom work that the same process won't hold up later.
That distinction matters even more in the AI era. Software can be built and copied quickly, and some startups hit impressive revenue milestones before they've figured out a durable way to acquire and keep customers. At TechCrunch Disrupt, Lovable go-to-market lead Luke Arno, Polsia founder Ben Broca, and Theory Ventures' Tomasz Tunguz talked about how founders can use their first 90 days of commercial activity to build a sturdier foundation.
The point isn't to delay selling. It's to treat early sales as data you should test - not as a green light to scale on autopilot. This article walks through how to spot the difference between scattered revenue and repeatable demand, what to learn from customers, and when adding a sales team is likely to help (instead of just adding cost).
Why early revenue can create a false signal
Revenue answers one basic question: did someone pay? It doesn't tell you why they paid, whether they'll keep paying, or whether you can find more customers like them without blowing up your costs.
That gap is easy to miss. A founder might land early deals through personal contacts, founder-led demos, one-off discounts, or bespoke product work. Each sale is real. But if every customer has a different problem and takes a different path to purchase, those wins often won't add up to a repeatable sales system.
This is the difference between commercial traction and repeatable demand. Traction means something worked for someone. Repeatable demand is the reason you can expect the same approach to work again with similar customers.
A sale is evidence, not a verdict
Early sales are useful because they force the product into the real world. A customer has a budget, an existing workflow, alternatives, and a reason to act now. Their decision can teach you more than a long list of hypothetical feature requests.
Still, you need to look at the conditions behind each deal. Ask:
- What problem was the customer trying to solve?
- What made the problem urgent right then?
- How did the customer hear about the product?
- What part of the buying process created confidence - or friction?
- Did the customer use the product after paying?
- What would make them renew or recommend it?
Compare those answers across customers. Don't treat them as isolated anecdotes. One customer might buy because the product saves time, another might buy because they're experimenting with AI, and a third might need custom integration work. That variety can be helpful during discovery - but it doesn't automatically identify a clear market or sales motion.
AI makes the distinction more important
AI has sped up how quickly teams can build software. That helps you test ideas, but it also makes it easier for competitors to copy what's easy to see. A product that pulls in attention today might struggle to keep customers once the novelty wears off or a similar tool shows up.
So the "durable advantage" isn't necessarily a feature nobody else can copy. It might be your understanding of a specific customer's workflow, your ability to solve a recurring problem well, and your distribution path to the people who actually share that problem.
That's also why headline growth needs context. Lovable has reported very rapid growth - more than $600 million in annual recurring revenue - and a valuation above $13 billion from its investors, according to the source report. Those numbers can happen in an exceptional case. They're not a sensible default plan for every new software company.
Tomasz Tunguz pointed out that expectations for execution speed have risen as more capital flows through the startup ecosystem. That pressure is real. But copying the visible pace of a standout company - without understanding its market, product, or distribution - isn't a strategy. The better question is simpler: "What would we need to see so the next customer acquisition looks more predictable than random?"
Use the first 90 days to find repeatability
Treat the first 90 days after a product is ready to sell as a learning period, with revenue attached. You're not trying to avoid growth. You're trying to find patterns before you commit to a bigger, more expensive go-to-market setup.
Luke Arno described it as building "learning density" and pattern recognition around where customers come from. Practically, that means running enough focused customer conversations and selling attempts to see which problems, buyers, and acquisition paths keep showing up. Doing "a lot of activity" isn't the same as learning. You can end up with more noise than clarity if the conversations don't share a common hypothesis.
Start with a narrow hypothesis
Before you schedule a pile of sales calls, write down a working hypothesis. It should spell out:
- Customer: A specific kind of person or organization - not "anyone who uses software."
- Problem: A recurring, costly, or time-sensitive job they need to get done.
- Trigger: The event that makes them look for a solution now.
- Value: The outcome they expect, in their language.
- Path to purchase: How they discover, evaluate, approve, and start using the product.
This isn't a permanent positioning statement. It's a test. If it's too broad, customer feedback won't be easy to compare. If it's too narrow, you might miss adjacent demand. Either way, the goal is to make your assumption visible so the team can update it as you learn.
Record what happened, not just what was said
After each sales conversation, write down the customer's context and behavior. At this stage, a lightweight shared document or simple CRM is plenty. You don't need a fully developed sales-ops setup before you know which information actually matters.
Track where the lead came from, the problem they described, their current workaround, the objections you heard, time to decision, how much customization they needed, and what happened after they purchased. Separate what they said from what they did. Enthusiasm on a call isn't the same as activation, ongoing usage, or renewal.
Then review the notes regularly. Look for repeated phrasing, common triggers, similar objections, and a consistent path from first contact to actual value. If multiple customers describe the same problem in their own words and follow a similar buying process, that's more convincing than a pile of compliments from unrelated people.
Treat different sales paths as diagnostic information
You might win ten customers in ten different ways. That can produce revenue and show you different use cases, but it doesn't automatically prove you can scale acquisition. Arno's point was that bringing in ten customers to solve ten different problems through ten different motions doesn't give you much confidence in a repeatable go-to-market process. Three customers acquired through a similar motion to solve a similar problem usually tells you more.
And no, the number itself isn't a magic threshold. Three matching deals doesn't prove product-market fit, and ten varied deals doesn't mean the company is doomed. The real question is consistency: the more repeatable the customer, problem, and buying path, the more informative each additional sale becomes.
Use mismatches to refine the hypothesis. If customers have the same problem but find you through different channels, the problem may be clear while distribution is still uncertain. If they come from the same channel but ask for unrelated features, your distribution might work while your product focus needs work. Treat these as separate problems - don't collapse them into a vague claim that "things are going well."
Build a go-to-market motion before a sales team
Hiring salespeople can increase capacity. It doesn't automatically create a sales process. If the founder can't explain who buys, why they buy, and what consistently moves a deal forward, a new hire will likely reproduce the founder's uncertainty - just at a higher cost.
A go-to-market motion isn't just a pitch deck or a list of leads. It's a practical sequence that connects a specific customer problem to a repeatable way of finding, converting, and serving buyers. It can be imperfect. It just needs to be clear enough to test and improve.
Separate the parts of the motion
When you review early deals, break the process into stages:
- Reach: Where did the prospect come from? Referral, outbound, community, content, or an existing relationship?
- Relevance: Did the prospect recognize the problem and treat it as important enough to address?
- Evaluation: What did they need to see before they trusted the product?
- Purchase: Who approved the spend, and what blocked the decision?
- Activation: Did the customer reach value without heavy founder involvement?
- Retention: Did usage continue, and what would make them stay?
A weak stage can break an otherwise promising pattern. Maybe the product solves a real problem but the approval process is long enough to make the unit economics painful for small deals. Or maybe customers sign quickly but never adopt the product into their regular workflow. The right response depends on where the system breaks.
Know what you are asking a hire to repeat
Sales hires make more sense when there's a defined job to hand off. You don't need a perfect playbook or a huge archive of historical data. But you do need some evidence that the founder's approach can be understood and replicated.
Before hiring, try to answer these in plain language:
- Which customer profile should the salesperson focus on first?
- Which problem should the first conversation center on?
- What signs suggest a strong (or weak) prospect?
- Which objections keep coming up, and what responses have worked?
- What does a successful first month of product usage look like?
- Which parts of selling still depend on founder judgment?
If your answers change drastically from deal to deal, the next investment may be more customer discovery, product work, or founder-led selling - not simply a larger sales team. If a pattern is emerging, hiring can help test whether it holds beyond the founder's network and personal involvement.
Scale the constraint, not the excitement
Waiting too long has costs. You can miss a market window, overload founders, or lose customers because the team can't keep up. But scaling too early also has costs: payroll, management overhead, and a sales process that may optimize for the wrong customer.
Follow the bottleneck. If qualified prospects are showing up faster than the team can serve them, added capacity might be justified. If the uncertainty is still who the product is for, more sales capacity won't fix that. If customers buy but don't use the product, you may need work on retention and onboarding before acquisition ramps up.
In the real world, the answer is rarely "wait until everything is certain." Certainty arrives too late. The better standard is whether the next investment tests a known constraint - or just assumes the uncertainty will disappear as you grow.
Stay close to customers as you grow
Finding a promising customer pattern isn't the end of discovery. Markets shift, products evolve, and customers often only reveal the real job after they try using the software in an actual workflow. The feedback loop that helped you find demand has to survive your first hires and your first real push for growth.
Ben Broca from Polsia emphasized staying close to customers through direct conversations, email, and events. His company uses AI agents to build and operate businesses with very little human staff; Broca said he was its only employee. That setup is unusual, but the underlying lesson generalizes: being near customers helps you notice when your assumptions stop matching reality.
Make customer contact a system
"Talk to customers" is good advice, but it's incomplete if the conversations don't lead to something useful and repeatable. Founders can set a regular cadence for interviews, onboarding reviews, and follow-ups with active customers and customers who churned. The schedule should match your product and customer base. The key is avoiding reliance on only the loud complainers or the people who praise you the most.
A practical feedback loop includes:
- A clear reason for each conversation: discovery, onboarding, adoption, renewal, or churn.
- Questions about recent behavior, not predictions about hypothetical features.
- A place to capture patterns and share them with product and engineering.
- A decision owner who can explain what will change - or why it won't.
- Follow-up that tells customers what happened with their feedback.
This isn't an invitation to build every feature customers request. Customers are good witnesses to their problems, but they're not always the best source for the best solution. Your job is to separate the underlying need from the proposed feature - and then weigh that need against what other customers are asking for.
Community is useful when it creates learning
Community can help you understand how customers talk about a problem, share workflows, and support each other. It can also become a distraction if you measure attendance or membership without checking whether the activity improves product use, retention, referrals, or actual learning.
The same goes for events and newsletters. Those can be worth it, but only if they connect to a real customer or business outcome. A smaller event that produces a clear set of conversations with the right buyers can teach more than a big audience with no real intention to purchase.
As the company grows, founders should preserve access to customer reality while spreading responsibility. Sales, support, product, and engineering each see different parts of the experience. A useful system pulls those observations together instead of leaving customer knowledge trapped in one founder's inbox.
Watch for hype fading
A startup can benefit from a wave of interest, especially when a new technology category is getting attention. But curiosity and lasting value aren't the same thing. Keep asking whether customers keep using the product after the first experiment, whether they'd notice if it disappeared, and whether it becomes part of a recurring workflow.
Retention isn't a substitute for acquisition, and acquisition isn't a substitute for retention. They're connected parts of the same system. If customers churn because the product doesn't deliver ongoing value, pushing more people into the top of the funnel usually just fills up the same leak faster.
Conclusion
Early revenue matters, but it's not the same thing as a scalable business. The first 90 days of selling should help a startup understand which customers share a real, meaningful problem, how those customers make buying decisions, and whether the product keeps delivering value after the initial sale.
Before you add sales capacity, look for a motion you can explain and test: a recognizable customer, a recurring problem, a credible path to purchase, and evidence of ongoing use. If those pieces still change deal to deal, keep learning. If the pattern is solidifying and demand is exceeding your ability to serve it, invest in the constraint.
A practical next step: review your last ten customer conversations and sales. Group them by customer type, problem, acquisition path, and post-purchase outcome. The pattern - or the lack of one - should point directly to what your next 90 days are for.
This article is based on discussion reported by TechCrunch at Disrupt.