Sifter

Garrett Diamond's journey to Sifter began with a stark contrast. His first job out of college exposed him to a well-oiled issue tracking process; his second job was the complete opposite. Later, while doing consulting and agency work, he discovered his biggest struggle: getting non-technical clients to actually use bug trackers. "We tried half a dozen different bug trackers and they'd log in once and be like, 'I don't know about this.' Then they'd never log in again and just email us issues." He realized the problem wasn't the trackers themselves—it was that they were built by developers for developers, with no consideration for non-technical users.

Garrett didn't initially plan to build a business. "I just started designing it and mocking it up, kind of thinking, 'Oh, maybe I'll create something open source.' I didn't want to deal with support and I didn't want to have to take money online because this was pre-Stripe." But by sharing his mockups publicly, he generated interest. "My first customer kind of got me," he recalls with a laugh. People encouraged him to turn it into a real product, and he eventually gave in.

He quit his job in January 2008 and launched publicly in December 2008, with the beta starting a month or two before. Total development time: about 6 months of focused effort (out of 11 months juggling freelancing and Sifter). He used Ruby on Rails and MySQL—nothing fancy. The initial product was intentionally minimal: no file uploading, no search. "Almost too simple, uncomfortably simple." He wasted about a month building a custom billing system (something that wouldn't be necessary today with Stripe). His philosophy was clear: "It doesn't matter what a bug tracker is capable of. If the team won't use it, it's useless."

Garrett's early customers came from the community that had watched him work in the open. By the time he launched, he already had validation and interest. But his target audience was unusual: not the developers paying for Sifter, but their non-technical clients. This made marketing tricky. "I was removing my customers' customers' pain. And that made it difficult to have a compelling message for them unless they had that problem and really understood it well."

Despite the crowded bug tracker market, Garrett felt confident in his differentiation. Most bug trackers were built by developers for developers. "I was more of an information architect interested in user experience," he explains. "I felt confident that there weren't going to be many other people that cared about doing that." He deliberately avoided looking at competitors: "I was familiar with lots of bug trackers from earlier experience, but once I started building Sifter, I couldn't have cared less."

Garrett's primary marketing tactics included blogging, public sharing, and targeted advertising. Over 8 years, he spent roughly $20,000 on ads across various channels. The standout win was a $2,500 Daring Fireball ad in early 2009 that brought in 30-35 customers—a huge win when he had only ~100 customers at the time. "That was the only thing I think that really moved the needle," he says.

But Garrett learned that there's no silver bullet. "There's rarely anything that will change the course," he reflects. "There were lots of little things that kind of adjusted the course." He experimented with AdWords but found it unfulfilling and hard to maintain consistent effort. Sponsorships and podcast ads worked but required sustained attention.

One major mistake Garrett identifies in hindsight: over-relying on email for customer feedback. "I spent so much time talking to people via email that I thought that was a good enough proxy for understanding our customers. But in hindsight, getting on the phone and talking to people was a totally different story." Phone calls revealed small pain points customers wouldn't bother emailing about, opportunities to fix things in "five seconds and do a release."

The business also faced real crises. Once, Garrett lost eight hours of customer data while resizing servers. He was panicked—"This is it. We're out of business. It's done."—but recovered about three hours of data and issued $700 in refunds (free months for affected users). Customers were surprisingly understanding. "When you're honest and transparent and really working to help your customers, people are human. They are forgiving."

Sifter grew steadily for 8 years, generating healthy monthly recurring revenue. Garrett never found a single explosive growth tactic, but the business was stable and sustainable. Then, personal circumstances changed. Over three years, Garrett dealt with serious leg and foot issues, eventually facing amputation. The recurring revenue model proved invaluable: "The recurring revenue was fantastic and made it very, very easy to focus on health issues because Sifter still maintained and went on and did just fine."

But juggling medical recovery with business demands exhausted him. After hearing Patrick McKenzie's positive experience selling Bingo Card Creator through FE International, Garrett decided to explore a sale. He sold Sifter for low six figures and joined WildBit, a company he'd long admired (Sifter had used their email service Postmark). He's since become a vocal advocate for discussing the "long, slow SaaS ramp of death"—the reality that recurring revenue businesses grow slowly and require years of consistent effort, not overnight success. Through his writing and interviews on Starting and Sustaining, Garrett now helps other founders understand what it really takes to build and maintain a SaaS business.