A Long, Ugly Year of Depression That’s Finally Fading

Yesterday morning I woke up early to speak at the Business of Software conference in Boston. It was my first time there, and it’s an exceptional group. Then, after some meetings, I spoke in the afternoon at Hubspot’s Inbound conference (thankfully just across a long skybridge that connects Boston’s World Trade Center from its Convention Center). This morning, I woke up even earlier and caught a plane to Denver, where I presented on a panel with Moz’s investor, Brad Feld as well as Ben Huh of Cheezburger and Bart Lorang of FullContact at Denver’s Startup Week. Tomorrow, I’ll be reprising my role of organizer for Foundry’s CEO Summit (despite no longer being CEO).

Right now, I’m supposed to be at a dinner for Denver Startup week, but I’m all people’d out. I need to be an introvert. And I need to get back to blogging. And I really need to tell this story.

dryteardrops
via Maria Louiza Biri on Behance

These past few days, I’ve met dozens of new people and talked to lots of old friends and colleagues, too. The exchanges always start the same: “How are you doing?” I get a pang of dread every time I hear it.

That’s because, when I’m asked, I try, most of the time, to be honest. I shrug my shoulders, look at the ground, and give a half-hearted smile alongside a verbal response like “not too bad,” or “could be worse,” or “I’ve been better,” or “eh, kinda sucky.” Compared to the 13 months between May of 2013 and June of 2014, I’m a lot better. I was seriously depressed for a long time, and a lot of people close to me knew it even when I didn’t know myself. I wrote about that some in my post Can’t Sleep; Caught in the Loop and I talked about it a little in my Dichotomy of 2013 post, too, but let’s call a spade a spade. Rand had depression.

Depression is why, in the last 18 months, I blogged the least I have since starting my career. It’s almost certainly part of why my engagement in all sorts of things has suffered. It’s why I could hardly bear to be away from Geraldine for more than a day or two at at time. And, it’s a big part of why I stepped down as Moz’s CEO.

That depression, I believe, stems from shame. I was and remain, ashamed of myself due to a series of wholly avoidable missteps I made at Moz. Brad likes to say that every company he and the Foundry team have ever invested in go through tough times and most have near-death experiences before they emerge into their mature form. Even with my darkest lenses on, I can’t say that Moz was ever “near-death” in the past couple years. But I can say that we – that I – made dumb move after dumb move, and when those moves revealed themselves to be dumb, I pushed us to double-down on them, compounding our problems. By sharing that story I hope to, as I’ve done in posts like this before, offload my pain through transparency and, hopefully, help save a few of you reading this post from making the same mistakes.

A Wrong Assumption

In 2011, Adam Feldstein, Moz’s Chief Product Officer, and I got together in a room and whiteboarded a series of goals and wireframes that we collectively dubbed “Moz Analytics.” It was to be a next-step in the evolution of Moz’s software suite, and a successor to our old “SEOmoz Pro” campaign product that tracked keyword rankings, crawl stats, links, and competitive data. When we sat down with Brad and Foundry in 2012, Moz Analytics was already underway, though we knew the project was understaffed given the seemingly massive task at hand. Our pitch was simple – over the next 5 years, we felt that marketers from many different backgrounds would be sharing the workload around efforts like SEO, social media, content marketing, online branding, competitive tracking, link building, etc. Basically – the disciplines that fall under “inbound marketing” and sit at the top of the customer acquisition funnel would all come together and be run by the same person (or people).

These marketers, we reasoned, would need a comprehensive suite in which they could manage all their SEO and social and content and brand marketing efforts in one place. We wanted to raise money to build that suite and grow our customer base beyond just “SEO” to these other disciplines, too.

My memory’s imperfect, but as I recall, in those early days, nearly everyone at Moz and at Foundry was on the same page – the idea made sense, and Moz was a company that could get this done.

The Worst Possible Way to Build Software

When the Foundry investment closed, we redoubled our efforts to build Moz Analytics. We hired more aggressively (and briefly had a $12,000 referral bonus for engineers that ended up bringing in mostly wrong kinds of candidates along with creating some internal culture issues), and spent months planning the fine details of the product. That product planning led to an immense series of wireframes and comps (visual designs of what the product would look like and how it would function) that numbered into the hundreds of screens. Our engineers estimated 6-9 months of work, with a launch date in July of 2012.

That July date came and went pretty fast. I was frustrated to find out that, based on our speed, October was more realistic, but I figured we had time, and we could accelerate some of innovation through acquisition. We spent a bit of our new VC money on Followerwonk and, later, GetListed (now Moz Local), along with AudienceWise.

Guess what? The date slipped again. And again. And again. From October to November to January (2013) to March (2013) to May (2013) and finally to October 2013 when the actual release occurred.

Many times, someone from our engineering team sat down with me during a 1:1 and told me that we were building software in the worst possible way, and that big bang releases are incredibly tricky to pull off, and always take much longer than expected, especially when you have an existing userbase to migrate. Anthony (our CTO) suggested numerous times that we instead build up to the features we eventually wanted on the existing SEOmoz Pro platform, but I felt that the marketing launch could never be as successful and couldn’t coincide with our rebrand – something that had been part of the plan all along. Moz Analytics and the shift from SEOmoz to Moz were always meant to go together I reasoned, and while delay hurt us, we were still growing, and tons of folks were still enjoying and getting value from our existing software. So long as we could release this amazing behemoth in the next 60 days (and I always thought we could until the last date push from May to October), I figured it was worth the wait.

What made me so foolish? Why was I so bullheaded? The answer’s obvious — I was arrogant. We had pulled off big releases in the past that everyone doubted – first with Linkscape (our web index, that no one believed we could build in 9 months on a $1mm budget), then with Open Site Explorer (which took 9 months as well, but only launched 1 month late), and then with SEOmoz Pro (which also launched a month late and with some serious bugginess for the first few months thereafter). I thought this would be another launch like those – a little late, but worth the wait.

In April of 2013, we had a very rough board meeting and a series of rough executive team discussions. We ended up replacing the engineering lead for Moz Analytics (who, humbly and awesomely, stayed with Moz another full year working as an individual contributor on the team) and after 2 weeks, Dudley, who took over, informed us that the earliest he thought the team could deliver the completed Moz Analytics would be in October of that year, moving out the expected timeframe from what we thought was 30-60 days away, to almost 6 more months. Worse still, we could only hit that date if we removed almost a dozen big features, including the entire “Content” section of the app – one of the most important, requested, and valuable pieces.

I could barely believe it, but had to accept reality. The big bang strategy had failed. The new, more iterative approach on the work to date would have to suffice. Building on top of the old codebase was no longer even an option. It was a shitty, shitty time, and when we announced the new schedule to the Moz team, I think morale at the company hit an all-time low.

Below is an email I sent to the team in May, just before we’d been scheduled to launch Moz Analytics and Moz.com together. It’s overly dramatic by a long shot (the risk of things going bad for the company was massively overstated, both in my own head and this email), but it may lend a sense for how I felt after more than a year of thinking launch was happening any day now (please also forgive the bad Lord of the Rings analogy, based on some prior emails about the launch)…

launch-is-moved-email

BTW – we never came close to having to worry about a layoffs situation, though even bringing up the word in an email (even just to say “we’re doing stuff to prevent having to worry about that”) created a lot of fear. It would take months of repeating “no layoffs” before I undid the harm that word in this email caused. Hopefully that can be one lesson you take away – “layoffs” is a Pandora’s Box-type word at a startup. Don’t use it unless you’re really being transparent (and not just fearful and overly panicked as I was).

The Rebrand that Worked

One of the decisions we made in May of 2013, after Dudley informed us of the new launch date for Moz Analytics, was to decouple the rebrand from the product launch. That was probably a smart move (and one we likely should have done much earlier). Our marketing and inbound engineering team (who handled the entire redesign of the website and the new signup funnel) executed brilliantly. I’ve still never seen a more exciting week for Moz than the week we announced the rebrand and started gathering signups for the new product. The positivity, joy, and outpouring of support from our community gave us all a renewed energy. I recall collecting the team all in one very full room (the lobby of the old Mozplex) and announcing metrics from the day’s work. Tens of thousands of new folks had already signed up to get beta access to Moz Analytics. Our site had its highest visit day ever. We received more press than we’ve ever gotten. I just looked – I had more than 350 messages of congratulations and outreach in my inbox in a 24 hour period.

Over the next few months, the rebrand continued to perform mostly positively. One slight downside was a slight loss in search traffic while the engines caught up with all the redirects, and another was how we’d chosen to manage our conversion funnel. Because we wanted our audience focused on the new product and not-too-attached to the old SEOmoz Pro, we first showed off Moz Analytics, collected an email address for the beta list signup and only then let folks sign up for the old product if they still wanted. That meant slightly less growth and slightly higher churn from those cohorts (likely because they anticipated something bigger and better to come).

At the end of September, more than 90,000 people had expressed interest in Moz Analytics. Even though we were tired, frustrated by our past mistakes, and reticent to get too excited, we felt a bit hopeful. Even with low conversion rates (we estimated as little as 5% of an opt-in list that specifically signed up for a beta product), it could be a huge bump for the business after a slow summer.

The Software Launch That Didn’t Work

When Moz Analytics finally did launch, it wasn’t pretty. Dudley and the engineers had told us to expect a buggy, rushed product, and unfortunately, that’s what happened. In addition the bugs, we got a lot of complaints about UI & UX, about missing features, and about the relatively low value add that all the new things we’d built into the platform provided. This wasn’t from everyone – there was plenty of praise too – but it was heavy enough that, in the months that followed, we’d struggle to stay at a flat trajectory (not where a startup that’s raised lots of VC money should be).

Below is a slide deck I shared at the last All-hands company meeting of my CEO tenure at Moz:

I thought it would be a transparent, authentic way to share how things were going at the time and the journey we had ahead. Unfortunately, it didn’t go over that way and I received a lot of concerned feedback, in particular from Matt Brown and Sarah Bird, that the tone was too negative, too harsh, and, while it might be representative of my truth, wasn’t representative of reality. The depression was playing a part in that, but I didn’t understand it, yet. That “Loop” that was playing in my head over and over every night, costing me sleep, eating at my happiness and resolve, was also warping reality. I couldn’t see the forest or the trees – just the swamp.

Distractions that Took Us Off Course

Moz Analytics wasn’t merely a hard, painful project that had a lot of early stumbling points, it was also a distraction from a lot of the important work we could have been doing instead to help our customers. The Mozbar languished. Open Site Explorer languished. Mozscape languished (despite a tremendous amount of effort that, hopefully, will be paying off soon and has been built much more iteratively).  Our other research tools languished. Without attention from our team, some of these products came to be eclipsed by competitors who focused exclusively on one particular piece of the products in our subscription. It was heartbreaking to hear over and over again how one of my SEO friends had stopped using some particular Moz tool because they’d found tool X instead to take its place.

Had we been listening to our customers, iterating on the projects and products that mattered to them, and not consuming all of our development time and energy on long-delayed, poorly launched megasuite that did lots of things they didn’t need, we’d have been in a much different place at the end of 2013.

Learning our Lessons

Today, 9.5 months into 2014, we’ve been making those investments, and they’ve paid off. August was our best month since June of 2013. Churn rate is down. Monthly signups are up. The new Mozbar is beloved. Fresh Web Explorer’s alerts have been a hit. Keyword Difficulty & SERPs Analysis got some upgrades.  The new features in Moz Analytics like better onboarding, estimates for keyword referral data and the new action dashboard have increased rates of usage, vesting, and overall customer satisfaction. Due to the nature of SaaS businesses, 2014 will be a rough-looking year. We’ll probably grow 10% or less in overall revenue. But 2015 looks promising, even to a skeptic like me. A recent analysis by Scott Krager on SERPs showed us comparing favorably on some metrics to Hubspot, who just filed to go public.

mozvhubspot

That said, I am far from satisfied. The content section that was supposed to come out with the first release of Moz Analytics still hasn’t launched – in fact, dev work on it still hasn’t begun (that’s how much catch-up work we’ve needed to do). Years of effort into a new version of the Mozscape index are still a quarter or two away. There’s a new Open Site Explorer coming soon, but it’s not out yet. But, I think our biggest lag has been one of leadership — and one that’s now rectified.

Not being CEO has some good things about it, but it has some big frustrations too. I’ll share those in a future post. I don’t believe we would have made this change when we did if it weren’t necessary. I really wish that the CEO change could have come at a different time – a time more like now rather than like how things were in January. But I think maybe it took new leadership to help us turn things around (and that’s very hard for me to say, because I desperately want to believe that I could have led us here, too).

A Hopeful, But Different Future

People say it takes a big person to admit mistakes, and admit that they don’t have what it takes to lead. I don’t feel very big. In fact, I feel like what my Dad always told me I was – a high potential, low achiever kinda kid. High potential – I could have made those right decisions, or at least not invested for so long when the writing was on the wall. But instead, I’ve caused us to underachieve against our possibilities. And despite that, Mozzers have continued to support me. Our community has continued to support me. Brad and Michelle have continued to support me.

In July of 2013, I held my performance review with Sarah. It’s a tradition we’ve had, to review one another, and in thinking about the past and writing this post, I dug through my inbox to see what big things were happening that frustrating summer. I feel like this review (which I’ve excerpted just a few bits from) is telling:

rand-perf-review-july-13

We hadn’t yet talked about a potential move for me out of the CEO role, but there, at the end of one of the sections, in my own writing, are the words “If this company is to survive and grow, it needs a better leader.”

I feel hopeful because today, we have a better leader. I feel hopeful because she’s not an outsider with a “great resume,” installed by investors, but someone who’s been through all the same pains and struggles and hard lessons-learned that the rest of us have, and unlike me, she doesn’t overly internalize or overly dramatize bad results. As she’s expressed to me many times – hard problems get her out of bed in the morning. She loves working on them. I wish that I could have been the person to get us through some of these hard problems, and make the road, especially the first 6 months of this year, a little less hard for Sarah.

But, most of all, more than our investors, more than myself or my family, more so than even the Moz team, I felt like I let down the tens of thousands of marketers who believed in Moz. In the depths of my depression, it wasn’t Foundry or Ignition or Sarah or any other Mozzer that I was hitting my head against wall over – it was our community. I feel like I made a promise that I didn’t live up to these last few years. I’m so, so sorry. I promise to try harder smarter.

Depression

Depressed Rand is weird. Don’t get me wrong, regular Rand is weird, too. But depressed Rand magnifies the bad 10X and minimizes the good. He refuses to even acknowledge good news and, because he’s a pretty smart guy, he can usually argue for why that good news is actually just temporary and will turn to shit any minute. The weird part is, I think depressed Rand is actually a very authentic version of myself. When I felt depressed, I upheld TAGFEE – particularly the values of transparency and authenticity – as the reasons why I could and should be such a raging, all-consuming, negative naysayer.

For months, I worked with CEO coaches, and then therapists to try and disconnect my personal happiness from Moz’s company performance. I thought that was a critical goal that I needed to do in order to get back to a better place. Eventually, this Spring, I gave it up. I decided that in order to be happy, Moz had to do better, and I put more of my mental and emotional energy there, rather than lamenting a rewiring I couldn’t make my head perform.

The moment of change happened after a doctor’s visit, in a very weird, inexplicable way. A few months prior to the doctor’s visit, I had tried a chocolate truffle laced with THC (don’t panic, personal Marijuana use is legal in Washington State). I’ve had sciatica in my left leg for years and if you’ve hung out with me, you’ve probably seen me stand up uncomfortably or use a scarf or sweater as lumbar support to sit down. For 6 hours after I tried the pot truffle, I felt horrible – crazy paranoid and high and awful. Then, for the next 3 days, my leg didn’t hurt at all. For the first time in six years. I told my doctor about this, and she said “the funny thing is, Marijuana doesn’t have any pain-killing properties. It just lessens tension, anxiety, and stress for some people.”

Bam. Moment of epiphany. My pain is tied to anxiety and stress (probably) and since I’m not particularly excited about trying pot truffles again after those first 6 hours of hell, I need to work on those things in order to fix my leg. And not just my leg. There’s likely all sorts of maladies, mental and physical, that I might suffer in years ahead if I can’t get this under control. That night, I felt different. I looked at our daily numbers email in the morning, and I didn’t come up with some reason in my head why they were bad even though they looked good. Like the corn in Hyperbole & a Half’s famous comic, suddenly everything wasn’t hopeless bullshit.

Don’t worry if this doesn’t make sense to you. It doesn’t make sense to me, either. I know I fucked up. I know I let down a lot of people. I know I was depressed. And I know that right now, I’m OK. I could be worse. Some days are still kinda sucky, but not mega-sucky anymore. The future feels like it’s back to being a possibility, rather than just an area of thought I need to avoid if I want to sleep tonight. In fact, I think I’m going to sleep pretty well tonight.