What I Would Do Differently If I Started AgencyHandy Today

The Slack message arrived on a Tuesday afternoon, between a support ticket about Stripe webhooks and a cold brew that had gone warm on the corner of the desk. A customer we’d spent four months building a white-label invoicing module for was cancelling. Not because the module was broken. Because they’d never opened it. “Honestly,” he wrote, “we just needed the client portal to not look embarrassing. The rest was noise.” I read it twice, then looked at the roadmap pinned to the wall — three quarters of which existed to please customers exactly like him.

That was the afternoon most of my SaaS founder lessons learned stopped being theoretical. Everything I’d done to feel safe — ship more, charge less, hire ahead, say yes to everyone — turned out to carry a bill. Nobody sends you that bill. You just pay it quietly, in the form of a product nobody asked for and a runway that’s shorter than your spreadsheet claims. Here is what I’d do differently, and more importantly, what the comfortable version of each decision actually costs.

Why do SaaS founders keep building features nobody uses?

Because building feels like progress and saying no feels like neglect. In 2019, Pendo analyzed 615 SaaS products and found that 80% of features are rarely or never used — 56% never touched at all, with only 12% used frequently.

That number sounds like someone else’s problem until you map it onto your own changelog. For a bootstrapped team, every unused feature is not a rounding error — it is a salaried engineer-month, a support surface that now needs documentation, and a settings page that scares off the next trial user. Pendo put the global tab at $29.5 billion in annual cloud R&D spent on features that go unused. I contributed my share. The white-label invoicing module, the bulk CSV importer three customers requested and zero adopted, the “advanced permissions” matrix that generated more confused tickets than it ever prevented.

The comfortable lie is that more features reduce churn. The boring truth is that the safe-looking roadmap — the one stuffed with “customers asked for it” line items — is the single most expensive thing in the building, and it never shows up as a cost. It shows up as velocity, which is how you fail to notice it.

Is it safer to launch a SaaS with more features or fewer?

Fewer, by a wide margin — the loaded launch is the riskier one, it just doesn’t feel that way. The Startup Genome report, after studying more than 3,200 high-growth startups, found that 74% failed from premature scaling: building product, team, and commitments faster than they validated demand.

The same report found that companies which scaled at the right pace grew roughly 20x faster than those that scaled prematurely. Read that again — the cautious-looking founders who shipped the “complete” platform weren’t de-risking. They were quietly choosing the slower-growth, higher-mortality path while telling investors they were being thorough.

The market does not reward you for the features you shipped. It bills you for them, monthly, whether or not anyone logs in.

If I started AgencyHandy today, the v1 would embarrass me a little. A clean client portal, order intake, and invoicing. That’s it. The version I actually shipped had a service catalog, team roles, a help-desk module, and a half-built CRM, because a feature-complete product felt responsible. It wasn’t responsible. It was four extra months of burn defending surface area no one had paid to validate.

Why do experienced founders still underprice their SaaS?

Because a low price is the most socially comfortable number in the building — nobody on a sales call ever pushes back on “too cheap.” The cost of that comfort is structural, and it compounds.

When you launch agency software at $29/month because $29 feels frictionless, you don’t just earn less. You select for the customer segment most likely to leave. SaaS revenue churn benchmarks in 2024 cluster around a 10–12% annual median, but that average hides a brutal split: the cheapest tiers churn at multiples of the premium ones, because price is a filter for commitment. I spent two years convinced our churn was a product problem. A good chunk of it was a pricing problem wearing a product-problem costume. The customers who paid the least complained the most, used the least, and left the fastest — and I’d hand-picked them with my own price page.

The hidden cost of the safe price is not the margin you leave on the table. It’s that you build your entire company — your support load, your roadmap pressure, your churn cohort — around the customers who value you least. Underpricing isn’t humble. It’s a decision to be busy and broke, made one reasonable-sounding discount at a time.

Should a SaaS founder hire to move faster before product-market fit?

No — early hiring is the premature-scaling trap with a friendly face, and it’s almost always sold to you as the responsible move. Every founder who hires ahead of validated demand is told they’re “building the team to execute.” What they’re actually doing is converting flexible cash into fixed obligation before they know what the obligation is for.

CB Insights’ analysis of startup post-mortems found that 42% of failed companies cited “no market need,” and a 2024 update of 431 failed venture-backed companies put poor product-market fit at the top at 43%. Notice what’s missing from the top of that list: “didn’t hire fast enough.” Nobody dies from a small team. They die from building the wrong thing efficiently with a big one.

I hired a second engineer and a part-time support person before I could prove our retention curve flattened. It felt like de-risking. It was the opposite — it raised my monthly burn by a number that turned every slow month into a crisis, which in turn pushed me toward shipping more features to justify the headcount, which is how the safe choice quietly finances the other safe choice. The comfortable decisions don’t just cost money individually. They subsidize each other.

Does it cost more to chase every customer or to say no?

Saying yes to everyone is far more expensive — it’s just that the bill is itemized as “growth” instead of “distraction.” Every agency that signed up wanted us to be a slightly different product: marketing agencies wanted social scheduling, dev shops wanted Git integrations, consultancies wanted retainers and time tracking.

The conventional wisdom is that a horizontal product serving everyone is the safe, large-market play. The data disagrees. The Startup Genome work is blunt about it: chasing scale and breadth before a repeatable, validated model is the dominant failure mode, not the safe one. Serving everyone meant our onboarding had to explain a tool that did a little of everything and the entirety of nothing. Trial-to-paid conversion is where you feel this — a generalist demo is a longer demo, and a longer demo converts worse.

A product built for everyone has to be re-explained to each person, and every sentence of explanation is a place the trial quietly dies.

If I started today, I’d pick one agency type and be almost rudely specific. The narrow product is the one that feels risky and is actually safe, because a customer who sees their exact workflow on the homepage doesn’t need to be sold — and doesn’t churn looking for the tool that was built for them specifically.

But isn’t playing it safe how a bootstrapped founder survives without funding?

This is the objection I’d have raised myself three years ago, and it’s wrong in a specific, measurable way. The argument assumes “safe” means “low-risk.” For an unfunded SaaS, the actual scarce resource is not money — it’s the number of decisions you can make before you run out of months. Every dollar spent on an unused feature, every customer acquired at a price that guarantees churn, every hire made before validation, spends a month you can’t get back.

Here’s the evidence that closes it. CB Insights consistently finds “ran out of cash” near the top of failure causes — cited by the majority of post-mortems — but the data underneath shows cash is the symptom. The companies didn’t run out of money because they were too bold. They ran out because they spread thin cash across the comfortable, conventional, diversified bets that each felt prudent in isolation. The “safe” portfolio of small reasonable choices is what empties the account. Concentration — fewer features, higher prices, one customer type, a small team — is the bootstrapper’s actual risk management, even though every individual concentrated choice feels like the dangerous one in the moment. Safety, done by instinct, is just risk you’ve agreed not to look at directly.

So here is the one prediction I’ll make as someone still inside this, not above it. Over the next two or three years, the agency-software market will quietly sort itself into narrow tools that own one workflow completely and bloated platforms slowly bleeding the customers they over-served. The bloated ones will look healthier for a while — more logos, more features, bigger decks. Then their churn cohorts will mature, and the bill for every comfortable choice will come due at once, the way it always does: not on a dramatic day, but on an ordinary Tuesday, in a Slack message from someone who only ever needed the client portal to not look embarrassing.

Sources

• Pendo, 2019 Feature Adoption Report — https://www.pendo.io/resources/the-2019-feature-adoption-report/
• Startup Genome, Why Startups Fail (Premature Scaling) — https://s3.amazonaws.com/startupcompass-public/StartupGenomeReport2_Why_Startups_Fail_v2.pdf
• CB Insights, The Top Reasons Startups Fail — https://www.cbinsights.com/research/report/startup-failure-reasons-top/
• SaaS churn benchmarks, 2024 — https://usermotion.com/saas-churn-rate-benchmark-2024