Why marketing software will never be like ERP

I ran across a thought-provoking article earlier this week on GigaOm — Marketing is the next big money sector in technology — by Ajay Agarwal of Bain Capital Ventures. It begins with the reference to the Gartner report from December the suggested by 2017, CMOs will have larger technology budgets than CIOs, which is a great place to start.

Ajay predicts that a new wave of companies leveraging big data for the benefit of driving marketing insights, “will create several multi-billion dollar winners. And a set of technology companies will emerge as the marketing equivalents of Salesforce and SAP.”

Hurrah!

He takes a moment to point at several of Bain’s investments in marketing technology ventures: BloomReach, CQuotient, HookLogic, and TellApart. (Hey, for all the benefits of big data marketing, a little PR is still in vogue.) Some very cool companies, actually. He then lays out two main points:

1. The multi-billion juggernauts of back-office enterprise software will have counterparts in the front-office worlds of sales and marketing (well, acknowledging that Salesforce.com, and Siebel before it, already achieved that status).
    
2. Much of the focus of marketing software has been focused on “process,” but there’s a huge new opportunity for a focus on “data” — and the web is an overflowing cornucopia of such data.

Here’s Ajay’s graphic of enterprise functions and some of their notable venture successes:

There’s conspicuous white space at the bottom, no?

Ajay remarks, “Despite the last 15 years of automation of sales functions, marketing functions have been underserved and underpenetrated in terms of enterprise software. While the other corporate functions have all created multibillion-dollar software companies, the marketing function has only had one exit north of one billion (Omniture).”

At a high level, I vigorously agree with Ajay on both counts: there’s a huge enterprise opportunity in marketing, and we’ve only begun to tap the potential of “big data” in this field.

But I think marketing may play out differently than those other enterprise functions before it. I’ll explain why in a moment.

The missing billions in enterprise marketing successes

First, however, I do feel that there are several arguments for multi-billion dollar marketing businesses that aren’t represented in the above chart. On the enterprise software angle, both Adobe and IBM have created huge ventures in the space — granted, by the aggregation of many acquisitions and good, old-fashioned organic growth. Teradata probably deserves a mention in this category too, especially with their acquisition of Aprimo. And ExactTarget is well on its way to a billion dollar IPO. Constant Contact has a near-billion market cap of $882 million.

There also an agency angle — I know, a different beast, but one that is increasingly engaged in coopetition with the likes of Adobe and IBM — Digitas was bought by Publicis for $1.3 billion. As for technology service providers focused on marketing, there’s Axciom ($1.1 billion market cap), Accenture ($40 billion market cap), Sapient ($1.7 billion market cap).

But the most interesting angle for claiming that marketing has had some spectacular multi-billion dollar successes: Google and Facebook. They’re not enterprise software providers, but the lion’s share of their multi-billion dollar revenues comes purely from selling technology-powered advertising-as-a-service to marketers. You could argue that they’re really media companies, but that’s a straw man. Google and Facebook are letting marketers leverage their “big data” and algorithmic prowess to achieve very intimate customer interactions in ways that have innovated marketing far more than any enterprise software package.

(Also note that as a piece of that puzzle, DoubleClick and aQuantive were both multi-billion acquisitions by Google and Microsoft respectively.)

My point is simply that the marketing function hasn’t been as anemic in big venture wins as that chart suggests. It’s just that the nature of marketing successes has been very different than those in other corporate functions.

Why the marketing software space is structurally different

One of the strongest indicators to me that the marketing software space is fundamentally different than finance, manufacturing, or HR, is the vibrant and varied ecosystem of marketing technology providers:

The above graphic — which, from 6 months ago is already woefully out of date — is just a sample of about 250 companies out of thousands. Many are small, venture-financed start-ups. But many aren’t so small either, representing revenues and valuations of tens or hundreds of millions of dollars each. In aggregate, it’s a multi-billion kaleidoscope.

There was never anything of this scale of diversity in software markets for other corporate functions. Why?

Two words: stability and standardization.

Finance, manufacturing, even the bulk of HR, may be complex functions, but for the most part, they’re highly standardized. In the case of finance for public companies, it’s downright regulated. GAAP is GAAP, no matter which finance software you’ve adopted. Getting creative with finance isn’t just discouraged, it can get you thrown in jail. So even though different software packages might have better ways of organizing, visualizing, or managing these standardized processes, the core of what they do is the same.

This meant that there were only so many companies worth funding in those spaces in the first place. And as they matured, it was relatively easy for the biggest players to swallow smaller competitors in consolidation. In many cases, those weren’t technology acquisitions — they were customer base acquisitions. They could even get away with discontinuing the acquired products and forcing customers over to their primary platform because they were essentially the same thing.

But not only are these functions relatively standardized — they’re also highly stable. With a few subtle differences, finance is managed today the same way it has been for decades. Same for manufacturing. The global scale of manufacturing has expanded dramatically, which increases the complexity of supply chain management to be sure, but the fundamentals are the same as they ever were.

In other words, software providers in these other functions aren’t frantically chasing a moving target. And that gives potential new entrants precious few opportunities upon which to mount an attack.

That’s not to say that there haven’t been a few big innovations. For instance, sales force automation and CRMs were pretty much locked up by the late 90’s — until Salesforce.com seized the opportunity to provide them as software-as-a-service. It’s still basically the same function, but delivered in a better way (less IT overhead). Manufacturing and other back-office departments may yet have their SaaS revolutions. But that won’t necessarily change the fundamentals of those functions.

Admittedly, HR may be a growing exception to this rule. Why? Because companies are starting to make radical changes in how they manage an increasingly mobile and global workforce. Shifting attitudes and expectations about work, epitomized through the Millennial generation, are opening up opportunities for HR software that embraces The Cluetrain Manifesto. But as important as these changes are culturally, from a technical perspective, I suspect there are finitely limited ways by which software providers may reinvent the technology that powers them.

There will be some amazing HR software start-ups, but there won’t be anywhere near the gargantuan surge we’ve seen in the marketing domain.

Stability and standardization do not characterize marketing today

The marketing software space could not be more structurally different than these other functions.

First and foremost, we must acknowledge: the entirety of marketing is in a perpetual state of disruptive innovation. Google changed everything. Then YouTube. Then Facebook. Then Twitter, Foursquare, Groupon, Quora. Apple has changed everything in mobile marketing and apps on our phones and tablets. My list of 131 different kinds of marketing offers a taste of the exponential explosion of new marketing approaches. And the pace of change only seems to be accelerating.

The modern marketing landscape is in constant flux: it is the antithesis of stability.

I’m not just talking about marketing software providers either. The very channels and vehicles they work on top of are still rapidly evolving, with new ones being invented every week. (Take Pinterest for example.) Like my examples of Google and Facebook above, you can revolutionize enterprise marketing without even being an enterprise marketing software provider.

It’s a breathtaking whirlwind — so much opportunity for competitive advantage. But it’s also dizzying and overwhelming. No one person, even no one organization, can keep on top of it all. Even trying to specialize in just one subfield, such as search engine marketing, is a high-speed roller coaster ride. Blink and you miss some huge development that changes the landscape.

This broadly disruptive environment is fertile ground for start-ups, who can see the potential of a new marketing mechanism and rise quickly with its success. Larger marketing software platforms seek to add support for these new mechanisms as features — sometimes just by acquiring one of those plucky start-ups — but they must continually battle inertia and bloat. The more add-ons you hurriedly slap on to your once sleek race car, the more it begins to maneuver like a Winnebago.

Until along comes a new platform provider that capitalizes on all the learning of how to connect the dots of those disparate marketing mechanisms into a better cohesive whole. Which is very exhilarating, of course, until the next series of innovations starts to drag their architecture in unexpected directions. And the cycle repeats.

But the dynamics of products vs. platforms vs. suites are something I’ll save for a future post.

The other axis by which marketing software differs from other functions is in standardization — that is to say, marketing software has almost no industry-wide standardization. Part of that is due to the lack of stability; it’s hard to standardize when the ground keeps shifting under your feet. (The only things that are standardized are the APIs that big channels like Google and Facebook enforce to a standard. And those standards evolve darn fast.) But there’s a deeper force working against standardization that emanates from the very soul of marketing: the Holy Grail of differentiation.

Marketers don’t want their customer experiences to be standardized against their competitors. They want their brands to be unique and beautiful snowflakes.

That drives a constant demand for creative new ways to stand out from the competition, “think different.” And the digital environment is infinitely malleable in this regard, far more easily plied by imagination than physical reality. The most common limitations? The architecture, features, and philosophy of the software you’ve adopted.

Because there is no GAAP equivalent standard in social media marketing — GASP (Generally Accepted Social Practices)? — for how to universally interface with customers on Facebook or Twitter, marketers and marketing software providers are free to experiment with a wide range of different approaches. You can have two diametrically opposed solutions that are both successful in different contexts. One that works great for one brand might be a train wreck for another.

So marketing software providers themselves are differentiated by the approaches they embody in their products, which guide and influence the strategies and tactics that marketers design and execute with them. As I wrote in a previous post, for marketers, you are the software you use.

But there’s one more way in which marketing software differs from other enterprise functions: process.

Process in marketing is simultaneously important and impossible

Let’s step back for a moment to acknowledge that life is process. Things move forward in organizations via processes — whether those processes are formal or informal, planned or improvised, optimal or dysfunctional. It may be a messy process, but there is one. Process is how strategy materializes, for better or worse.

And a huge portion of the value of enterprise software for other corporate functions has been the systemization of processes it enables and enforces — the grand-daddy of which is enterprise resource planning (ERP), the broad umbrella of Oracle and SAP.

See, while human beings struggle with process on their own — remember all the exact steps to take, at the right time, in the right order, in coordination with everyone else involved in the process — software is ideal for such precision synchronization. That was a big part of the value that software in these other enterprise functions was able to deliver. Automated and semi-automated processes can be incredibly efficient.

But here’s the key difference: to a very high degree, businesses can control their processes for finance, manufacturing, and HR. You can bend all of the participants to your will. “This is the process. Follow it.” Employees, distributors, vendors, contractors, subcontractors — if they want their checks, they bow to the process. Wal-Mart’s supply chain is the epitome of this for an extended enterprise.

It’s worth nothing that, along these lines, sales force automation was about automating salespeople, not sales. Important distinction, and confusing the two can lead to delusions of automation. (“Press this button and sales just flow in!”)

Marketing is made up of processes too. But the problem, like sales, is that there are two very different worlds of process that marketing faces:

1. Processes within the marketing department (and its various collaborators).
2. Processes that prospects and customers follow to engage with the business.

While processes within the marketing department are as scriptable as any other enterprise function — e.g., here are the steps you must follow to execute an approved campaign-specific special offer — you can’t dictate process to your prospects and customers. Or, more accurately, you can try to dictate process to them, but they can choose not to follow it. After all, your customers are primarily beholden to the processes of their own organizations. When there’s a conflict, their process wins, not yours.

And this is why marketing software will never be like ERP — we can’t dictate the process end-to-end.

Marketing is — or should be — obsessed with their customers’ processes. How do we optimize the experiences we deliver to flow smoothly and successfully with their processes? If we’re going to suggest that they follow our processes, at least in some circumstances, how do we make that enticing and rewarding? Since there’s almost always tremendous variation among customers, this is very tricky. One size doesn’t fit all.

You could take some comfort in the the notion that at least processes internal to the marketing department should be able to be systematized by software. And some can be. But here’s the rub: even internal marketing processes become entangled in addressing customer processes. Many of marketing’s processes are inherently about acting on or reacting to customers.

This is why marketing automation is so damn difficult.

It’s not that marketing automation software sucks. It’s that its mission is impossible to achieve perfectly. The best we can do is systematize pieces of our marketing ecosystem, carefully interweaving what we automate, what we semi-automate, and what we leave under manual control.

The balance is never perfect — but we can always asymptotically improve it. Of course, with the caveat that because marketing is neither stable nor standardized, so that balance is constantly changing. But when we get it even approximately right, the payoff can be huge — which is why I say that process in marketing is both important and impossible.

Which leads me back to Ajay’s article, where he proposes that a new generation of marketing software can be more focused on data than process — as a way of overcoming the limitations of the process-driven approach of current marketing automation solutions.

I do agree that more data can help us make better decisions in a number of marketing processes. Yet I still think that process design and management will be the overarching challenge for marketers (and the software platforms that serve them). Better data and smarter use of that data will enable major advancements in marketing’s capabilities. But I’m skeptical of the ability for data to self-generate process — at least outside of very specific contexts. Data is potential; process is activation of potential.

So software that takes a hybrid approach to process and big data strikes me as the most promising.

Still, no solution will be perfect. Machine learning, as powerful as it is, is a statistical method — there’s always a margin of error. And if you apply machine learning to the wrong data, or extrapolate its conclusions into unjustified territory, or even just foul up the presentation layer of what it outputs, you can end up with a big flopping sound.

Modern marketing software won’t be perfect, but it will be amazing

I have no doubt that the years ahead will have several billion-dollar exits in enterprise marketing software. The mostly likely candidates will be platforms that can serve as a backbone for core marketing processes and data repositories, your marketing “systems of record.” Arguably, Salesforce.com is one example of this, since the CRM is increasingly owned by marketing. Adobe, Eloqua, IBM, Aprimo, Hubspot, IgnitionOne, Neolane, Pardot, ExactTarget, and many others are all competing to be the skeletal structure of marketing’s IT. New entrants will arrive for sure.

But unlike other enterprise functions, where a single platform such as SAP or Oracle pretty much covers everything, I don’t believe that backbone marketing systems will be complete one-stop shops for the entirety of the marketing department.

Instead, I think such primary systems will be complemented by a wide variety of so-called “point solutions.” Almost all of them will interface to backbone systems, but their specialization will be the key to addressing the instability challenge — and differentiation dream — of marketing for the foreseeable future. (The big backbone winners will be those who make such interactions with other software especially easy.)

The particular combination of backbone systems and point solutions will vary dramatically from company to company, depending on context, industry, vision, brand. The unique fabric of any one particular company’s marketing software mix will be a big part of their competitive differentiation.

As a result, this ecosystem will continue to support a tremendous number of smaller ventures, with many more waves of new marketing software start-ups every year. These will result in a number of exits at tens of millions or hundreds of millions of dollars. In aggregate, these will represent multiple billion-dollar subfields within marketing.

Rather than fantasize of “one platform to rule them all,” marketers, entrepreneurs, and venture capitalists should embrace the complex, heterogeneous environment of modern marketing. There’s no precedent for it in the history of enterprise software. But frankly, there’s no precedent for any of the amazing revolutions that are erupting in marketing writ large.

It’s also a great time for marketing technologists and a move towards agile marketing management. But I’ll close this post here.

Agree? Disagree? Share your perspective in the comments.