Thinking through the recent Andreessen-Horowitz investment in Mixpanel has got me thinking about the overall economics of the digital analytics industry over the past few days. (Plus, my wife and I finished all of The Americans, so unless I get back to re-watching Battlestar Galactica over the holiday before season 4 of GoT hits iTunes, there’s nothing good on TV.)
As I’ve said before, I believe that digital analytics is a foundational business tool for any modern enterprise, and what we’re seeing in almost every industry today is a divergence between those companies that have evolved to embed this technology into their strategy and execute on it, and those that have not. Digital analytics – web, mobile, and the like – enables and complements all sorts of marketing and ecommerce technologies that are now table stakes for a competitive business, which puts the analytics vendor itself in an incredibly valuable strategic position. No wonder, then, that competition is so heated.
I’m going to take stock of the digital analytics vendor landscape as it stands at the beginning of 2015, make a few predictions about where it’s heading, and then circle back to the a16z-Mixpanel thing, because it’s a real head-scratcher. And it makes sense to start by talking about the two biggest names in enterprise digital marketing – Adobe and Google. In that order.
Adobe: seeing green
Love ‘em or hate ‘em, you have to give respect where it’s due. Over the last several years, Adobe has been one of the most innovative forces in digital marketing, and has executed on its strategy in a way that very few companies can. Adobe literally built a new pillar of its company with Adobe Marketing Cloud (which has since been adopted as the industry standard nomenclature for many of its competitors), and the business segment now accounts for about a third of the company’s overall revenue – a figure which is increasing. Adobe is now widely recognized as a leader in the industry, and rightly so.
Another reason I start with Adobe, however, is because unlike pretty much any of its competitors, the company is required to make public its financial results (for this business). My analysis of those results (click for a link to my Google spreadsheet analyzing their financial data) lead me to one conclusion: Adobe’s strategy is working.
By any measure, Adobe has been successful in building the Marketing Cloud business. Since 2010, Adobe has reported about $5.29 billion in revenue from their Digital Marketing segment (AKA “Omniture/Enterprise” segments in 2010/2011), compared to my estimated total of $3.1 billion combined costs in acquisitions. Adobe Marketing Cloud itself has an impressive 14.2% CAGR for revenue since 2011.
[Update, October 2015: I’ve updated my financial model for AMC.]
One interesting quirk about Adobe Marketing Cloud is that not all of its products are, of course, cloud-based. In fact, Marketing Cloud revenue is just over half SaaS (subscription) based. The other near-half, Experience Manager and Campaign (Neolane) businesses, are mostly on-premise, meaning that you can strip out those segments from AMC total revenue to get a rough estimate of the growth of the Analytics, Test & Target and Social components. That business, too, as illustrated in the chart above, has grown at a nearly 20% CAGR since Adobe began breaking out Digital Marketing subscription revenue in FY2012.
(Quick caveat here: since “cost of revenue” is a notoriously squishy category with no clear definition, I don’t mean to suggest that the difference between it and total revenue is Adobe’s profit margin here. Net contribution from AMC is assuredly much lower. Nevertheless, it is a key indicator.)
Adobe’s Marketing Cloud SaaS business is hitting its stride, and is steadily cleaving apart its Revenue from Cost of Revenue (as you can see in the curves above). This is a case study in economy of scale, and a hard-won long-term advantage for Adobe when selling against its enterprise-class competitors. With its extensive investments in client adoption, partner ecosystem and marketing outreach, Adobe is both able to accelerate upsell/cross-sell of its SaaS components to Analytics clients and position itself as a preferred vendor. As it gets better at that, Adobe can (and does) price more aggressively to win ever greater market share, displacing smaller “boutique” analytics vendors. (More on that later.)
Oh, and by the way: I predict that Adobe will announce about $387M in Digital Marketing revenue for Q4 (that’s the unaudited figures, which have not been released yet), with a 55/45 split between subscription and traditional license revenue. Revenue for the company as a whole is off 6% since 2012, and net income is down almost 70%, yet their P/E is through the roof (currently at 147). Crazytown.
So in case anyone is still wondering: yes, Adobe has a nice business going selling Analytics – because, of course, it’s not just about selling analytics. That’s what so many people don’t really get.
The Big G
First – in Q3 2014 alone, the cost of revenue for Adobe Marketing Cloud was $117 million. (As I said before, that’s not comprehensive costs, but probably accounts for most of the technical infrastructure.) They’re on track to rack up almost half a billion in costs this year. And Google Analytics is far, far larger than Adobe Marketing Cloud.
This isn’t an apples-to-apples comparison, of course – Adobe incurs much higher costs because its Marketing Cloud offers far more functionality, data retention, services, etc. And Google enjoys a distributed global datacenter infrastructure that spreads its costs across many of its lucrative business lines. But GA also services tens of millions of websites with increasingly sophisticated analytics, and does so for free. So when you really think about the scale Google operates on – and the costs they must incur to support GA – it’s really pretty mind boggling.
The other thing to understand about GA is that Google is just not playing the same game as the other analytics vendors out there. Google being Google, they have rigorously measured the impact that using GA has on the advertising spend by companies who use it, and found that there’s a positive relationship. I doubt Google will ever make that analysis public (I’ve only heard rumors about it), but when you consider the scale of their enormous investment in GA, you can begin to imagine what that ad revenue payoff must look like. It also explains the impressive integration between GA and Google’s advertising products.
Google Analytics is the standard by which all other analytics tools are measured, and for good reason. With every launch of new features or integrations, no matter how minor, the entire digital marketing world goes up in a tizzy. By doing so, Google also forces other (paid) vendors to constantly defend why their solution is better than “free,” and provides a major barrier to new entrants. To boot, their product team is top-tier and deeply connected in the #measure and marketing community. (Hey guys!)
All of that is formidable enough. But what I think is most interesting about Google’s strategy right now is what they’re doing with GA Premium.
Google Analytics Premium represents a meaningful prong in Google’s overall focus on enterprise software, and a major threat to incumbents like Adobe. With GA Premium, Google offers companies investing in new analytics capabilities the prospect of not having to re-platform on a new solution (sort of – upgrading to Premium probably does actually require some re-tagging), and end users can remain within an application everyone already knows well. Premium also allows Google to focus its product investment on a more lucrative market segment while building its brand cred for upselling additional Google for Work products as well. (This is what terrifies Microsoft.)
Google’s first big weakness is in all of the surrounding marketing technology offerings that the “marketing clouds” offer, and where they arguably offer their fullest value – testing and optimization, automation, customer experience, email marketing, and the like. I don’t see Google ever becoming a “marketing cloud” per se, but I would bet that they’re not done acquiring. Optimizely, for example, was founded by two Googlers and would fit neatly in the GAP stack. Google Ventures is a big investor in HubSpot (a detail not lost to Scott Brinker). And, of course, Google building an email marketing tool in-house that connected directly to Gmail would probably kill off a few competing ESPs overnight.
Its second big weakness is in its client services operation. While the GACP program is expansive, Google arguably has less control over the network of consultants who can actually do the services work necessary for its clients than Adobe does with its partners. While Adobe farms out a substantial amount of services work to its partner network, it also captures quite a bit of revenue by doing that work itself too. Many clients, particularly larger ones, demand that direct vendor support for their enterprise software investments, and services operations are not only difficult and expensive to scale, but Google is fairly allergic to them. I think the strength of the GA Premium product, Google’s brand allure and shifting market demands will help them peel off some customers from Adobe, but that their richest source of opportunity is surely in the hundreds of companies who are just discovering each year that new investments in analytics makes business sense. (Indeed, the Google Analytics marketing and education teams are instrumental in pushing that message.)
In any case, I think that the strength and scale of Google Analytics has contributed notably to the consolidation of “marketing clouds” all by itself, as other vendors race to both win digital analytics market share and differentiate by offering functionality on top of it.
Feeling the squeeze
AT Internet and Webtrekk have managed to carve out continued success, particularly in Europe (specifically the French and German markets respectively). As home-grown solutions in markets that have a strong preference towards local vendors, service and language support, both companies have effectively fought competition from the big players up until now. In a sign that those advantages are waning, however, both companies are now focusing on Latin America and Asia/Pacific markets for continued growth. Doing so while maintaining the lean organizations that made them successful (and, likely, profitable) will not be easy, however. Compensating for inevitably eroding market share at home (lost to Google and Adobe) will require winning a lot of business in one of those growth markets. Earlier this year, Webtrekk closed a financing round to do exactly this, and it has surprised me to hear nothing similar from AT Internet.
That leaves IBM and Webtrends. Yes, both are still kicking. I won’t say a whole lot here about the former (still a little too close to home!), but the latter is interesting. Earlier this year, Webtrends brought in the ex-CEO of Coremetrics, Joe Davis, to lead their organization and try to regain momentum in this market, and he’s made some big changes to the company. 2015 will tell if Davis is successful. It’s long been rumored that Webtrends was looking for an acquirer, and Davis, after all, brokered the sale of Coremetrics to IBM. So… who do you think would want a digital analytics capability? I can think of several emergent “marketing clouds” that need one.
Way over on the opposite side of the digital analytics food chain, you have several digital analytics startups that, while they offer interesting technology, have not (in my view) yet materially affected the major vendors above. With every new free feature launch from Google, and every price point drop for Adobe Analytics, these vendors are pressured from above and below for a raison d’etre:
In the long run, of course, most of these companies will not survive. Such is the nature of the startup game. Those that do manage to begin selling consistently into the mid-market (which very few of these do) will either have to heavily specialize (like Parse.ly), pivot entirely, or find that translating that success into enterprise accounts is difficult as long as they remain primarily suppliers of data – or even “insight.” “Connecting analytics with marketing action” is both a marketing tagline and a product development epic, and is precisely what is driving the successful consolidation of “marketing clouds.”
What is Mixpanel up to with Andreessen-Horowitz?
Let me say: I have a ton of respect for the folks over at both Mixpanel and Andreessen-Horowitz. Mixpanel is an awesome product which I’ve gotten to play around with, albeit one that I’ve never actually seen in a competitive deal. And I’ve gotten the feeling that those guys over at a16z seem to know a thing or two about the internet and stuff as well.
Suhail Doshi, Mixpanel’s CEO and a certifiable genius, says that his company has been consistently profitable for several years, and they plan to use the $65M raised from A-H, on a company valuation of $865 million, to fund acquisitions and R&D. For a sense of how things have changed, remember that Adobe acquired Omniture in 2009 for a bit more than twice as much – $1.8 billion – when it had 1,200 employees and about $300M in revenue, which utterly dwarfs Mixpanel today. No one knows exactly what Mixpanel’s revenue stream is like, but my guess is that valuing it at $865 million would require a very significant multiplier.
The kinds of value claims that Marc Andreessen makes about Mixpanel (for example, in his tweetstorm announcing the investment) are pretty much straight out of every analytics vendor’s marketing brochure in 2009: Mixpanel “makes state-of-the-art mobile/web analytics easy for every company in the world;” “Mixpanel lets you track actions not just pageviews,” and so forth. None of this is exactly revolutionary. Rather, the heart of the matter is here:
9/Mixpanel will go straight after the goal of predicting the future with data — what we think is the next phase of analytics.
— Marc Andreessen (@pmarca) December 18, 2014
This “next phase of analytics” is, in a way, similar to what IBM is working on with Watson (though clearly applied differently here), and to what Salesforce is talking about with its Wave product (which is still mostly just demos at this point). The holy grail, of course, is something like a predictive analytics tool that can parse any number of disparate datasets and respond to non-technical user queries with contextually relevant answers. Needless to say, that is a very, very big project to tackle, even for a company with deep development resources like Google, IBM, Adobe or Salesforce.
Mixpanel is probably hoping that it can short-circuit much of the “big company development process” and bring to market a robust such “future of analytics” product before Adobe does. Given the difficulty in dislodging incumbents platforms in this market, especially in larger companies, only a truly outstanding offering is likely to be successful.
Compounding the problem, though, is the fact that the wide fragmentation of marketing data sources (web, mobile, email, social, in-store, merchandising, ecommerce, CRM, etc.) creates an automatic advantage for those vendors who already host that data on their own platform – like the “marketing clouds.” Mixpanel has none of that other data in-house already, and will have to rely on third party connections/integrations to make any such “future of analytics” secret sauce work.
Could they do it, though? Sure. And I can’t wait to see what they come up with.
Digital analytics has been called a “commodity,” which it is not. The analytics space continues to evolve rapidly, but the differentiation and value proposition is increasingly moving up the stack from that foundation. The digital analytics foundation will be a necessary, but not sufficient, component of building a sustainable business in this market. Highly differentiated point solutions will continue to grow, particularly in the mid-market, but clients will demand integration with the primary analytics dashboards they increasingly use to gauge overall performance.
More than anything, though, what I see as the biggest continuing pains are onboarding of marketing technology, enabling its productive use, and integration between tools. Fear and uncertainty about how this is all done is holding back a lot of companies from increased investment, and for those that have invested already, technical hangups slow down the adoption curve substantially. Indeed, issues like these have created a whole burgeoning marketing technology services and consulting industry, where reputable providers have to turn away business because of too much demand. I don’t expect any change there in the near term – indeed, I wouldn’t be surprised to see a business services juggernaut like Deloitte or Accenture make acquisitions in that area.
So: fragmentation, fear and uncertainty, rapid evolution – good times all around! It’s a terrific time to be in ecommerce and marketing technology. This post has been way too long, so I’m going to stop here. But I look forward to seeing your thoughts or predictions in the comments below.
I’ve since written this update on Adobe’s results over the course of 2015. Check it out!