Summarize this article with:
The 27th client signed on a Friday. I remember because I’d promised myself I’d take the weekend off, and instead I spent it rebuilding our onboarding doc, because I’d just realized we had no real onboarding doc. Three clients to thirty in twenty-two months. No salesperson on payroll. No agency built on the back of a closer who could talk anyone into a retainer. Just me, mostly, and a system I didn’t fully understand was a system until much later.
I keep getting asked how. So this is the actual version, including the parts that were luck and the parts that broke.
We didn’t replace the sales team. We replaced the sales activity.
The phrase ‘without a sales team’ is a flex on LinkedIn and I’m guilty of using it. But it’s also misleading. We didn’t skip selling. We just moved the selling out of conversations and into artifacts. The pitch happened before the call. By the time someone booked a discovery slot, they’d already read three case studies, watched a campaign teardown, and seen us rebuild a Performance Max account on a public Loom.
The result was that the call wasn’t a sales call. It was a fit check. They’d already decided.
That distinction matters. Most agency owners reading ‘no sales team’ hear ‘no selling.’ That’s not what happened. We sold constantly. We just stopped doing it one person at a time.
The first 3 clients came from Upwork. Nobody likes hearing that.
The origin nobody wants to hear: our first three clients came from Upwork. Two were lead-gen accounts, one was a Shopify store running about $40K a month in ad spend. None of them came from a fancy outbound system or a content moat. I bid on jobs at midnight, wrote proposals that were too long, and undercharged because I was scared.
The Upwork clients did one thing that mattered later: they let me keep their data and case study rights. I asked for that in writing in month two of the first retainer. That single contract clause is probably the highest-leverage move I made in year one.
If you’re reading this from inside the bid-on-Upwork-and-pray phase, two pieces of advice. First, get case study rights early, in writing, before you’ve delivered a single result. Once they’re winning, they become protective. Second, don’t wait until you have ‘real’ clients to start documenting. The unsexy ones build the asset that gets you the sexy ones.
Six months trying to ‘do real marketing’ got us zero clients
In month nine, I decided we needed to ‘do real marketing’ for ourselves. We’d been winging it. Time to be intentional. So I built a content calendar, hired a part-time writer, started a podcast, ran lead-gen ads to a downloadable PPC checklist, and posted on LinkedIn five days a week.
Six months. Zero clients from any of it. A few hundred email subscribers. Some podcast downloads. A pile of MQLs from the lead magnet that did exactly nothing.
The mistake was treating agency marketing like client marketing. We were running funnels. Funnels work for products. They mostly don’t work for retainers above a certain size, because nobody buys a $5K-a-month relationship from a checklist download. They buy it from proof, and from watching someone think.
The six months wasn’t entirely wasted. The content calendar got destroyed. The podcast got paused. The lead magnet got buried. But the LinkedIn habit stayed. That part eventually paid for everything.
The pivot: every client became proof, every proof became a magnet
The pivot, when it came, wasn’t a strategy meeting. It was a single decision: stop publishing opinions, start publishing receipts.
Every client engagement got two outputs from then on. The deliverables they paid for, and an internal write-up I’d later turn into something public. Numbers, screenshots with the brand redacted if they wanted, the actual decisions we made, and the things we got wrong. Especially the things we got wrong. Those got more reach than the wins.
This sounds like content marketing. It isn’t. Content marketing is ’10 ways to lower your CPA.’ This was ‘Here’s the dumbest thing I did with this account in Q2 and why we recovered.’ The difference matters because the second one is unfakeable. Anyone in PPC can spot a generic tip article. Almost nobody can fake the texture of a specific, slightly embarrassing case.
That decision didn’t immediately bring clients. But six weeks in, I started getting cold inbound from people who’d read three or four pieces and were already pre-sold.
How the case study factory actually worked
The mechanics, because this is the part agencies usually skip when they tell this kind of story.
Every client got a write-up at month three and another at month nine. Two questions structured them. What did we change? What happened because of it? That was it. We resisted the urge to add narrative arcs and ‘key takeaways.’ The numbers carried the weight.
Each case study went three places. The website, gated behind nothing. A LinkedIn post, written as a story not a summary, usually 280 to 400 words. And a short Loom video walking through the actual account. The Loom was the highest-converting asset we had, by a wide margin. It made everything we said impossible to fake.
By month eighteen we had fourteen case studies live. Roughly half were Shopify e-commerce, the rest were lead-gen across home services, B2B SaaS, and one driving school client we still talk about. The variety helped more than I expected. Prospects didn’t need to see their exact industry. They needed to see we could think clearly about a different industry, which is what they were really hiring for.
LinkedIn wasn’t thought leadership. It was just receipts.
LinkedIn became the spine. Not because I love LinkedIn. Because every other channel needed an audience to start with, and LinkedIn was the one place I had ten followers who actually worked at companies that could afford a retainer.
The format that worked wasn’t thought leadership. Thought leadership is ‘Here’s what I think about iOS 17 and attribution.’ That post gets likes from other agency owners. It doesn’t get clients. The format that worked was ‘Here’s something I did this month, here are the numbers, here’s what I learned.’ Boring. Specific. Slightly vulnerable.
Roughly 40 percent of new business in year two came directly from LinkedIn DMs. Another 25 percent came from referrals where the referrer mentioned, unprompted, that they’d been reading my posts. So call it 60 percent attributable to one channel I wasn’t really trying to optimize.
If you take one thing from this section, it’s the difference between posting opinions and posting evidence. Evidence compounds. Opinions evaporate.
What 30 clients did to a 4-person team
Hitting 30 active retainers with four people on the team broke things.
Specifically, it broke me. By month 22, I was the bottleneck on every account that had any complexity, the only person doing sales, the person QAing the work of two of our three contractors, and the person writing the LinkedIn posts. I slept badly for about six weeks before I admitted what was happening.
The fix wasn’t more hiring, though we did hire. The fix was that I had to stop being the gatekeeper on quality and start being the architect of the system that produced quality without me. Those are different jobs. The first one feels productive. The second one feels slow and uncomfortable. Another contributor on this site wrote about killing the project manager role and what that cost them. Their version of the same lesson lands from a different angle.
We documented every recurring task that took me more than 20 minutes. We built loose checklists for the things that resisted SOPs. We promoted our most senior contractor to a salaried lead role, and gave them final sign-off on three of our largest accounts. By month 26 I was no longer the QA layer on most accounts.
The retention math nobody tells you about
Here’s the part the LinkedIn flex posts skip. Of the 30 clients we landed in those 22 months, we lost 7 within the first 18 months of their retainer.
Some left because we got results and they brought it in-house. Some left because we got results and they wanted aggressive scaling we couldn’t operationally support yet. Some left because we’d taken on accounts that weren’t a good fit, and we knew it during onboarding, and we said yes anyway because growth felt good.
Net retainer revenue was up. Logo count was misleading. If you’d looked at our 30 clients and assumed 30 retainers compounding, you’d have been wrong by about 23 percent on the revenue side.
The lesson, which took me too long to learn: client count is a vanity metric the moment it gets ahead of your delivery infrastructure. The number that matters is how many of your last 10 closes are still active 12 months later. That’s the only growth number I look at now.
What I’d do differently next time
If I were starting again with the same goal, three changes.
I’d skip the six-month ‘real marketing’ detour entirely and go straight to publishing receipts in month one, even if the receipts were small. Bad early case studies are still better than good early opinion content.
I’d say no to the wrong-fit clients in months 14 to 18. Saying yes felt like growth at the time. It cost us delivery quality and contributed to most of the churn that came later.
I’d hire the operations person before the second account manager. We hired in the wrong order and paid for it. The work didn’t need more hands. It needed more system.
The ‘no sales team’ thing wasn’t a strategy. It was a constraint that forced us to build something more durable than a sales team. That’s the only part I’d keep.
If you take one thing from any of this, it’s this. The agencies that grow without a sales team aren’t the ones that figured out a clever inbound trick. They’re the ones who turned every piece of client work into proof, and every piece of proof into something a stranger could find. That’s slow. It compounds. It doesn’t always feel like marketing while you’re doing it.
There are more essays from agency owners on the kind of growth that doesn’t look like growth at the time. If you’ve lived through one of these stretches yourself, the inbox is open.
About the author
Ishant Sharma is the founder of Hustle Marketers, a Google Ads and Meta Ads agency working with e-commerce and lead-gen brands across the US, UK, UAE, and Australia. Twelve years in performance marketing. Writes occasionally on agency growth and the unsexy mechanics of running a PPC shop.