I decided to create a tidy table view of Barrel’s business performance for the past 14 years cleaning up some numbers and simplifying the view to focus on revenue, cost of goods sold (COGS), gross profit, and gross margin. I decided to keep owner’s compensation completely out and defined COGS as the sum of our payroll and any amounts paid to freelancers/consultants/other agencies to help us deliver work.

One year immediately jumped out to me: 2015.

That year, revenues decreased year-over-year for the first time in company history, by 3%. Gross profits were down by nearly 22%. This, in the 14-year view, seemed to be when we hit “rock bottom.”

I dove in and pulled up some primary documents to see what had happened. I read old emails, my journal entries from the time, Evernote entries, spreadsheets, Harvest time-tracking records, Basecamp threads, monthly team meeting presentations, and whatever else I could find. It’s easy in hindsight to point at my older self and make fun of how clueless I was, but to my credit, I could see that I, along with some others at the company, were really trying to make things work.

What’s clear and obvious now wasn’t so straightforward back then. But I’m also sure that five years from now, the 2025 version of myself will look at my 2020 version and shake my head at what I couldn’t see and understand.

For my own records and also in case there are others who may be in or were in a similar situation, I’ve pulled out a few lessons gleaned from my 20/20 hindsight view.

Lack of Focus Running the Agency

Nearly a decade into running Barrel, I was still convinced that our destiny was to develop and launch software products that would one day surpass our client income. With profits we made from doing client work, much of it was plowed into “internal” efforts with the dream of building high-margin software businesses.

I was caught up in overseeing the design, build, and marketing of these products (see the end of this post for a list of the product graveyard). A few of them launched and generated revenues, but we faced some obstacles:

• We never quite made enough profit to fully fund these product teams. The team members would have to split time between doing client work and doing product work, often finding themselves working late hours to keep up. The result was subpar outcomes on both client projects and the products. We were continually plagued with buggy work and delays.
• The product teams were often run by me rather than a dedicated product manager. This resulted in chaos as I rarely put in place timelines, changed roadmaps on a whim, and forced people to pay attention to non-critical details that I deemed important.
• I had no concept of cost and did very little financial modeling to understand whether or not any of these products were viable as business units. I wrongly assumed that a great, well-designed product would easily pay for itself. The lack of financial insight also made me lose sight of the actual cost in building these products. Every hour that I took people away from billable client work meant we would have to make it up with additional hiring or freelancers. Unbelievably, I never took the time to really calculate these costs.

It’s hard to run an agency well even with 100% focus. Given my divided attention, it’s easy to see in hindsight that I was doing a terrible job of running Barrel’s client business and also struggling to oversee the products.

The end of 2014 was when I started to get the sense that Barrel as a company was on shaky ground. I found an old email to the team explaining the suspension of a SaaS product that I had devoted a better part of a year trying to get to market. This product had cost us hundreds of thousands of dollars in labor but was still buggy and lacking much interest from potential customers. I was also starting to lose interest in the dream of shifting to a software business. I began to wonder what would happen if I focused wholly on making Barrel a successful agency.

We finally sunset all of our non-client efforts in 2016, and it would take a painful 2015 for me to finally get the idea into my head: focus on running the agency.

Lack of Organizational Structure

While the team numbered close to 30 people, we were still a very flat organization. We had made distinctions between junior, mid-level, and senior levels for individual contributors, but a management layer was pretty non-existent. My co-founder Sei-Wook and I divvied up management duties across the entire organization, which meant we were responsible for the oversight of our most experienced developer as we were for our greenest design intern.

In having so many direct reports, we ended up not managing much at all. This created a very hands-off “figure it out yourself” environment where a bunch of young twenty-somethings would try their best to get things done and occasionally get a passive aggressive email from one of the founders if they made a mistake. It’s no wonder we began to experience quite a bit of turnover in the years leading up to 2015. In 2015 alone, we turned over 45% of the team as a large cohort of long-time employees were burnt out and bounced. Those that came in to replace them were quickly thrown into a “figure it out yourself” environment with a fair number of them leaving within 12 months.

I can’t say for certain why we didn’t invest in building out a proper organizational structure earlier. I think fear was a big part of it. The fear of hiring someone more experienced who would cost a lot more than our junior talent and potentially not working out. The fear of someone from the outside coming in and saying we weren’t doing things the right way. The fear of giving up control of all the decisions within the organization.

In researching for this post, I came across an org chart from 2015 that was laughably flat and under-developed. It was clear that we were really behind when it came to thinking about our employees’ long-term career development and the formation of discipline-specific cultures. Notes from this time do highlight the fact that we were aware of this challenge but it would take time and many missteps before we established something more permanent.

Today, it’s clear to me that a strong agency is one that is thoughtfully structured with well-defined leadership across all its functions. We’ve made a lot of progress in building out discipline leadership with supporting layers of management. There are still some functions that Sei-Wook and I still oversee, but our roadmap is clear: invest in developing and hiring really capable people who can take the reins and do a better job than we ever can.

Lack of Project Profitability

In 2015, we did a poor job of making money on our projects.

In reviewing some of these projects, I was able to pull out some common mistakes:

• The projects were poorly scoped and under-estimated. In many instances, we said “yes” to whatever number the client threw at us and proceeded to pour as much time as we felt was necessary to deliver a “quality” product.
• Projects dragged on for too long due to various delays, some caused by our team and some by the client. In both cases, the budgets stayed the same while team members stayed on and poured in more hours. We rarely paused to allow the team to work on other projects or issue a change order to get paid more.
• We eagerly embraced new technologies and ate the extra hours they would take to learn and implement on client projects, often leading to complications and additional debugging that caused project delays.

Here’s a screenshot from Harvest of one of several “projects gone wrong” from 2015. This particular website was severely under-scoped, dragged on for too long, and included some new technologies and platforms that we weren’t familiar with. The result? An effective billable rate of $38/hour for a project that took a significant chunk of time from some of our most talented people. I also learned that being wildly unprofitable on a project for the sake of putting out something “great”–where the client is happy and we win both awards and admiration from our peers–is ill-advised. Do it enough times and you may rack up the awards and respect but find yourself out of business.

The biggest shame is not that some projects were unprofitable in 2015, but the fact that we didn’t seem to learn from the experience and kept on doing more work unprofitably. Project profitability, believe it or not, was not a metric we kept back then. We were naively confident that there were enough profitable projects to make up for these money losers.

It’s been a multi-year effort to really get a handle on project profitability. Even today, I think we have a great deal of room to improve. We’ve also made progress on improving our scoping and overall project management, but we still have the occasional projects that miss the mark on profitability. I wrote a detailed post on the importance of calculating project profitability, but it’s one thing to share some sample scenarios and quite another to ensure the business can consistently execute. Thankfully, we’ve come a long way since 2015.

Lack of a Strong New Business Pipeline

In looking at our new business logs from 2015, a couple things jumped out to me: the quality of our leads were subpar and we didn’t put an emphasis on developing existing client relationships.

When it came to new leads, I could tell that we spent way too much time trying to land anyone who would pay us. We were terrible at saying no to poor fit clients in the off chance they could pay even a little bit for a website. This type of desperation happens when cashflow is weak from diminishing profits.

A large chunk of my nights and weekends were devoted to writing proposals for deals that we had no business in going after. Some were obvious poor fits in terms of the type of work or the budgets being too small. For others, we had no advantage and were pitted against a dozen other agencies, so it was a crapshoot at best. Even then, we pumped out proposal after proposal, hoping to land new deals to keep the business alive.

When it came to existing clients, we lacked a robust account management approach. Even with clients who were impressed by our work and happy with the results, we rarely took the time to learn more about what was happening with their business and explore ways to help. We waited until they reached out with a need and did our best to deliver, but we missed out on opportunities to surface pain points that we could help solve or brainstorm creative approaches to new campaigns and business objectives.

I would argue that we still haven’t quite cracked this challenge almost 5 years later. We’ve built stronger relationships with some of our clients and have become more strategic thought partners, but this is still relatively nascent. We are also not totally well-positioned or marketing in a way that gets us highly qualified leads. We get a good volume of leads but most of those are a poor fit. If anything, we’ve reduced spent time on proposals as we take far less swings. The percentage of revenue coming from existing clients has continued to grow, which is a good sign. As with most things mentioned here, we’ve made a lot of progress but we still have much work to do.

Systems Thinking and Discipline

The thing with business performance is that it’s the cumulation of dozens of decisions made over a long timespan. There are internal and external systems at play where decisions made today can have beneficial or disastrous impact weeks or months from today.

With the benefit of experience, we’ve become more and more aware of different systems and how our decisions impact them. This has allowed us to avoid problems or if we become embroiled in a problem, to at least understand why things are going wrong. An example is how we’ve come to view decisions made during the new business process and the downstream impact it has on finances, resourcing, positioning, and team development.

When looking back at 2015 and the years leading up to it, it’s clear that many of the systems that are clear to us now were hidden to us back then. As a result, a big chunk of my time was spent playing defense, always reacting to different fires on a daily basis. I can remember the anxiety of wondering every Sunday night what surprises were in store for me the upcoming week and how I would deal with it. It was only during rare vacations and holiday season when I could steal an hour or so to think about the company’s long-term future.

One of the primary benefits of seeing the business as an interplay of systems is the ability to set and test different rules and see how they perform. In other words, we’re able to go about running the business in a more scientific manner via hypotheses and experiments vs. trying to run things by feel or pure intuition. This is where discipline comes in. We can set parameters on types of new business we want to take on with a long-term view and be disciplined about saying no. We can set metrics on gross margins and be disciplined about hiring full-time employees only if the work can support it. We can give a candidate a thumbs up only if they’ve passed certain tests and went through a predetermined process. The way to navigate around systems we can’t totally control is to develop systems that are absolutely under our control and to keep on tweaking them relentlessly.

When I look back, this is the major difference. We believe in the power of systems and the job is not to fight fires but to continue designing and testing systems.

---

 

The Product Graveyard

Software products built by Barrel that never quite made it.

• LaunchEffect: A WordPress theme that launched in 2011 to much fanfare including being featured on TechCrunch and numerous other tech/startup publications. We featured a free and premium version and generated tens of thousands of dollars. Sadly, upkeep and providing support to all kinds of WordPress users turned out to be incredibly expensive. We sunset this product in 2016.
• Seasons, Weekend, and Mosaic Shopify Themes: We were early to the scene with Shopify themes when the e-commerce platform was still a startup. We had some really great 5-figure months of selling $180 themes to small shop owners but like LaunchEffect, the on-going support and upkeep of updating the themes were a killer. We sunset Seasons and sold Weekend and Mosaic to a company that specializes in Shopify themes.
• ProjectFlow: Our first software-as-a-service (SaaS) product with recurring revenue, ProjectFlow was basically a stripped down version of a Trello board with way less features. This experience taught me so much and yet if you read my blog post from 2013 talking about lessons learned from it, I clearly did not learn enough–it would be another few years before we pulled the plug.
• TeamInsights: This was probably the costliest internal project we ever had and it never launched publicly. TeamInsights was a performance management tool for small teams that was meant to make 360 reviews easier and to make project specific feedback more automated. A large chunk of 2014 was spent on TeamInsights, which I know had downstream impact on how the company performed in 2015.
• DesignCub: This was yet another distraction–a WordPress portfolio theme aimed at designers and agencies. As we learned of the headaches in supporting WordPress customers from LaunchEffect, we decided to pull the plug only a week after the product launched and recorded a single $35 sale.