Moz Dumps Amazon Web Services, Citing Expense and ‘Lacking’ Service

[Updated, 1/31/14, 12:01 pm] Seattle marketing technology company Moz had a worse-than-expected 2013 in terms of profitability and products. But what really jumped out at me in the privately held company’s startlingly frank review of the year was new CEO Sarah Bird’s blunt criticism of Amazon Web Services (AWS), which she says the company is leaving for reasons of cost, product stability, and service.

“We create a lot of our own data at Moz, and it takes a lot of computing power,” Bird says in her 2013 Year in Review post. “Over the years, we’ve spent many small fortunes at Amazon Web Services. It was killing our margins and adding to product instability. Adding insult to injury, we’ve found the service… lacking.”

An AWS representative declined to comment.

Moz had a tumultuous year. CEO Rand Fishkin gave up the top executive job. The company changed its name in May from SEOmoz, reflecting an Internet marketing landscape in which the practice of search engine optimization is merging with the broader category of inbound marketing. Revenues fell short of expectations.

The company’s inbound marketing product, Moz Analytics, “launched” over a rocky nine-month period that’s just wrapping up. The company generated gross revenue of $29.3 million last year, up 33 percent, which Bird described as “off-plan performance.” Sales had doubled in each of the prior two years.

She attributes the slower growth to delays, bugs, and and a lack of certain features as Moz Analytics was rolled out. The new offering was designed to upgrade a previous product with a loyal following. Bottom line, Moz had a $5.7 million loss in 2013 before interest, taxes, depreciation, and amortization. It didn’t provide a similar figure last year, but had been profitable.

“We knew we were going to burn in 2013,” Bird writes. “That’s why we took the $18 [million] Series B” funding round in 2012, led by Brad Feld’s Foundry Group with participation from earlier investor Ignition Partners. “This is a bigger loss than we planned on, though, and we’re disappointed we missed our goals.”

Bird says Moz has adequate cash to continue growing and expects to return to profitability by the third quarter.

You should read Bird’s entire post if you’re following the company or the inbound marketing business closely. (Bird, by the way, was the company’s chief operating officer until earlier this year when she formally took over as CEO from co-founder Rand Fishkin. Fishkin announced the transition in December and said 2013 had been harder on him personally than almost any year in the company’s history, mainly “due to the challenges of scale.”)

Bird says part of Moz’ plan to return to profitability includes a transition away from AWS to its own private cloud. That strategy goes against the narrative we’ve been hearing for the last few years about startups, who have built businesses on commodity computing power rented on a pay-as-you-go basis from big cloud vendors like AWS. The idea is that the public cloud model is cheaper, more efficient, and less risky than buying and maintaining private computing infrastructure. Lately, even big enterprises have gotten into the act, running more of their operations on public clouds from Amazon, Microsoft, and other providers. Moz is just a single example, but it’s one backed up with the kind of financial details we don’t often see from privately held startups.

Moz spent almost $2.4 million on cloud services—the “vast majority” on AWS—in 2011, or 21 percent of that year’s revenue; $6.5 million in 2012, 30 percent of revenue; and $7.2 million last year, 25 percent of revenue.

That got the company’s attention, so it embarked on a plan in 2012 to build its own cloud in datacenters in Virginia, Washington, and Texas.

“This was a big bet with over $4 million in capital lease obligations on the line, and the good news is that it’s starting to pay off,” Bird says. “On a cash basis, we spent $6.2 million at Amazon Web Services, and a mere $2.8 million on our own data centers. The business impact is profound. We’re spending less and have improved reliability and efficiency.

“Our gross profit margin had eroded to ~64%, and as of December, it’s approaching 74%. We’re shooting for 80+%, and I know we’ll get there in 2014.”

[Update, 1/31/13, 12:01 p.m.: See our latest post for more nuance on Moz’ decision from CTO Anthony Skinner.]

Time will tell if Moz is the exception that proves the public cloud rule, or an early indicator that public clouds aren’t the best fit for everyone. Given Moz’ penchant for transparency, it’s reasonable to hope we’ll hear more from them about this transition.

Trending on Xconomy

By posting a comment, you agree to our terms and conditions.

10 responses to “Moz Dumps Amazon Web Services, Citing Expense and ‘Lacking’ Service”

  1. Noah Yetter says:

    This will not end well for them.

    AWS looks expensive on paper because A) you’re paying for flexibility and B) it includes a lot of things you have to buy separately if you use a colo facility or get for yourself if you run a datacenter (like power and internet access).

    The fact that they managed to spend MILLIONS a year on AWS tells me they’re not even *attempting* to use it properly. Are they buying reservations? Are they shutting off unused instances? Are they matching instance sizes to workloads or buying the big ones Just Because? Are they leveraging spot instances for scalable compute-heavy workloads?

    Private cloud sounds great to the naive, because in theory it means you get the flexibility of virtualization and managed services combined with the lower unit cost of owning your own hardware. What that doesn’t account for is developing the expertise, staffing, budgeting, etc. that’s necessary to run your own datacenter. The whole point of cloud computing is to let Amazon, Microsoft, Google, do that for you because they already specialize in it. If your business is X, where X isn’t “running datacenters”, then you shouldn’t be running a datacenter.

    • DaeDae says:

      Shouldn’t they just use AWS for backup/overage for spikes, instead of running everything on it?

      Even with Reserved instances the price break doesn’t equate to the cost savings of running your own servers. Plus the premium on AWS bandwidth costs. I get it that the flexibility and ease of AWS is awesome, but if your business is struggling its better to live lean imho.

    • HuntingLamers says:

      What I am interested knowing is how the f you have managed to reach the conclusions you have stated based on this article. I hope nobody is expecting you to reach any mission critical conclusions at work. the advices you give CEO’s CTO’s and engineers over one simple talk back… I can only imagine what you can do with real factual data…..

    • Xtraordinary Hosting says:

      Noah. But it’s never as simple as black and white is it? There’s plenty of room for different shades of opinion and mixed solutions. If you have intensive constant CPU crunching needs then owning a few racks of virtualised x86 kit could provide SUBSTANTIAL savings compared to AWS. There is a bit of a myth that Amazon likes to perpetuate that they are cheap. Flexible yes, have huge scale yes, but not always best bang for buck.

      You say private cloud doesn’t give you the management skills layer, but AWS doesn’t provide much of that either. Running your own ‘datacentre’ (or using third parties to help you do it) can make perfect financial sense for some people. It depends on your needs and skills so don’t jump to conclusions without the facts. Take a more neutral view or find someone who can advise you better.

      It sounds like Moz have taken a considered view on this and it will end well for them.

  2. Chris Mack says:

    They’re not the first to do this. Zynga moved from AWS to a hybrid model and saw huge costs savings. AWS is a fantastic service for variable need computing. Once you reach a certain scale – it absolutely makes sense to use your own servers.

  3. udkl1298 says:

    Crawling on AWS is a no-no from what I’ve heard … AWS is awesomely suitable for other things though !

  4. Idon't Know says:

    AWS is unreliable, insecure, and has poor service.

  5. Dwagswpb says:

    Gmail goes down for 8 hours and I don’t here about firings at google. That kind of down time would never be accepted with a private cloud option. That is an example how our society accepts lower levels of service for cheap.

  6. Penny Stocks Guide says:

    Add LKEN on TOP of your watchlist, it’s going to be so big this week , +500% is expected

  7. dlsniper says:

    “Time will tell if Moz is the exception that proves the public cloud rule, or an early indicator that public clouds aren’t the best fit for everyone.”

    This made me laugh sooo hard! You speak about technology yet you never heard / experienced that there are truly no silver bullets that match for ALL users / use cases? Is it a real shock that a server costing $200 can buy you (read you can rent) 4 servers, each 4-5 times more powerful, on a dedicated hosting company? Need I say more?