Marketing Automation does not exist

David Bessis
Oct 10 · 14 min read

“Marketing Automation” does solve key problems for B2C organizations. But it won’t solve their strategic customer marketing challenges.

In the “Mad Men” era, marketing was 99% art and 1% science. The battle was about branding, positioning and messaging. But when you look at the user interface of a modern Marketing Automation solution, it’s all about diagrams, decision rules and workflows.

What happened? Did marketing suddenly become a branch of engineering? Have human judgment, storytelling skills and business acumen really been made redundant?

In a nutshell: B2C Marketing Automation is a misnomer. For B2C companies, what people call “Marketing Automation” is actually about Sales Automation and Journey Automation. Both use cases bring massive value and justify investing in “Marketing Automation” solutions.
But Customer Marketing is much broader than that. Despite all investments in automation, over 90% of B2C messaging traffic still consists of newsletters, promotions, product launches and similar communications. This is Real Marketing. Its mission is to generate new demand while helping with the dynamic priorities of the business. It requires human judgment, storytelling and business acumen.
It’s about time to acknowledge that Real Marketing is mission-critical. In fact, it continues to run the show. While it cannot be fully automated, the good news is that marketers have new technologies at their disposal to modernize Real Marketing and drive customer-centricity, productivity and great business outcomes.

1. What B2C “Marketing Automation” is really about

While B2B companies use Marketing Automation for lead generation and nurturing, B2C companies mostly apply it to interact with their existing customers. Messages sent by B2C companies to their customers break down into three categories:

  1. Messages triggered by a recent customer action: cart recovery emails, recommendations based on a recent visit, cross-sell opportunities after a purchase, and so on. These tactics are the primary use cases for B2C Marketing Automation and they are incredibly effective. But they can only react to recent customer behavior and thus can’t generate new demand. In this respect, they aren’t true marketing activities; they’re sales-oriented activities that aim to maximize conversion. In the digital world, they compensate for the absence of a physical salesperson. This is Sales Automation.
  2. Lifecycle messages: “are you satisfied with your purchase?”, “happy birthday, we have a surprise for you :)”, “your trip starts next week, are you ready?” You cannot build this without Marketing Automation. But again this isn’t core marketing. These nurturing and service-oriented activities aim to optimize the customer experience. This is Journey Automation.
  3. Everything else: newsletters, promotions, product launches, flash sales alerts, sponsored campaigns, inspirational content, and so on. These activities aim at generating new demand from existing customers and, in this respect, they are the Real Customer Marketing. But even when they are scheduled within “Marketing Automation” solutions, they do not really benefit from automation. Humans stay in the loop, typically on a weekly basis, as new decisions, new inputs and new content are always required. Real Marketing has fallen through the cracks of automation.

Don’t get me wrong. Sales Automation and Journey Automation can bring massive value and buying a “Marketing Automation” solution is generally a good business decision. Deploying it was a big effort, but you’ve done it. The conversion rates of your digital channels have finally gone up. Congratulations!

But beware what comes next.

By lumping together activities that have nothing to do with marketing under the big umbrella name of Marketing Automation, marketers have started cutting the branch on which they are sitting.

Why? Because Sales Automation beats Real Marketing flat out on all KPIs and conversion metrics.

And guess what. It’s even worse if your organization relies on last-click attribution. This shouldn’t be a surprise. Based on last-touch attribution, supermarkets would fire everyone but the cashiers, since they’re the only ones generating revenue.

In fact, most attribution models are skewed toward Sales Automation. Of course, there is a catch:

Sales Automation only makes money because it converts existing demand.

Once you’ve deployed Sales Automation and tweaked it over a few quarters, it offers no further room for growth. There are only so many trigger scenarios that make sense. The same applies to Journey Automation. Quickly you’ll hit an automation ceiling.

The only way forward is to drive more demand into the system. To do this, marketers have only two options: spend more on acquisition (which is extremely costly) or improve Real Customer Marketing (which is much more cost-effective.)

But if your organization is intoxicated with the misleadingly favorable metrics of Sales Automation, promoting Real Marketing will be a tough sell.

Accurately measuring the true business value of Real Marketing is hard. Long-term control group studies do demonstrate a massive impact on revenue, engagement and lifetime value. Unfortunately, these studies are costly and difficult to deploy. And they don’t even factor in that, by contrast with Sales Automation that only pushes items related to what a customer was already interested in, Real Marketing enables you to steer your customers towards the topics you choose. Revenue on items with excess inventory has more business value than revenue on items you would have sold anyway. When campaigns actually respond to strategic needs of your organization, their business value cannot be captured by attribution metrics.

Real Marketing is the neglected workhorse of B2C Customer Marketing. It looks old-fashioned and spammy, yet it is absolutely mission-critical. A testament that B2C organizations do recognize this (albeit reluctantly) is that Real Marketing continues to constitute between 75-98% of their Customer Marketing messaging volume:

Despite its bad reputation, Real Marketing continues to run the show.

At the end of this article, we’ll provide an explanation for the bad reputation: the biggest problem with Real Marketing is that it is insufficiently data-driven. The good news is that this problem can be fixed. Real Marketing can be made into the strategic growth machine it deserves to be. But for organizations to fix their chronic underinvestment in Real Marketing, they must first recognize the unique mission and the operational specificities that prevent Real Marketing from being fully automated.

How not to frame the Sales Automation vs Real Marketing debate

When you compare apples and oranges, the optics can be horrible.

This picture doesn’t show that there may be little room to grow automated marketing. Nor does it explain that non-automated messages actually create demand that feeds automated campaigns (who then excel at capturing the revenue attribution.)

But show it to your CEO and she’ll say: “Scale down the non-automated marketing. Fire the creative team. Double down on automation. Hire more engineers. The battle is over. Algorithms have won. Humans have lost.”

And then you’ll enter the marketing death spiral.

It won’t be pretty.

In 18 months you’ll be batch-and-blasting “30% off” coupons in a desperate attempt to drive demand.

How to frame the Sales Automation vs Real Marketing debate

The conversation with your CEO will be much more productive if you convey that sales and marketing are distinct activities, with distinct missions. Show her this picture and she’ll immediately get it:

Everybody knows that success happens when sales and marketing work hand-in-hand. Sales is often the last touch point (and captures attribution) but it makes more money when marketing drives more demand into it.

Neglecting Real Marketing cannot be the way to grow your business. Quite to the contrary. Here is a little diagram that can help you deliver the message:

2. Real Marketing is hard: stories from the trenches

Clarifying the respective roles of Sales Automation and Real Marketing is helpful. But what explains that Real Marketing didn’t benefit much from automation?

Over the past few months, I’ve had dozens of conversations with marketing executives at leading B2C organizations in the United States and Europe. Their stories shed fascinating light on the real-world challenges that drive but also constrain customer marketing campaign plans. Here is a selection:

  • “We have trigger campaigns and we have batch campaigns. A better name for the batch campaigns? Well, you might as well call them Marketing. Let it be clear: I’m not going to wait for my customers to come. I’m paid to drive demand and I have specific offers to push.” (US travel company.)
  • “Last Spring it rained a lot, and we ended up not selling our bathing suits. Who’s supposed to fix the excess inventory? Me.” (European apparel brand.)
  • “One of our partners is paying us to promote a niche destination. Who should we target? And how does it fit within our existing campaign plan and our competing commercial priorities?” (Leading global airline.)
  • “We’ve tried to add more automation, but except for lower funnel conversion tactics it didn’t really fit in. Robomarketing isn’t for us. As a luxury company, we must start from the creative.” (US luxury brand.)
  • “The problem with automation is that it doesn’t understand storytelling.” (European jewellery brand.)
  • “We want to be customer-centric, but at the same time our entire organization is product-led. How do I resolve this contradiction? That’s my #1 customer marketing problem.” (US retail company.)
  • “Will AI ever replace marketers? Honestly I don’t think so. Unless the AI can sit at our executive committee meetings, when we decide our marketing priorities.” (European apparel brand.)
  • “We’ve just created a single customer view that unifies customer data across all our entities & all our countries. Now everyone will be tapping the same database. All messages will be competing for voice-share. But we have no idea on how to solve the prioritization and governance problem.” (US media and entertainment conglomerate.)

While truly diverse, these stories share four striking common features:

  1. Marketing teams do not operate in isolation. They have to respond to incoming demands from their organizations. For all the dreams of customer-centricity, marketers are paid to help with the business.
  2. These demands are diverse and they change all the time. There are many stakeholders and external factors. Stuff happens. This week’s urgencies pile up onto last week’s priorities. It’s a dynamic problem, not something you can program once and for all on “autoplay.”
  3. It’s not about isolated campaigns. It’s about continuously building and rebuilding the whole plan, prioritizing competing goals and resolving conflicts. At a higher level, it’s about governance and strategy.
  4. Marketers may sit on tons of big data, but key data points are missing from the systems. What was said at the executive committee? How strategic is this partnership? How good is this new offer? Modern marketers continue to need strong business acumen and situational awareness.

3. Driving in Chandler vs Driving in Manhattan

These four features are the reason why Real Marketing is hard to automate. A comparison with self-driving car helps understand why.

Despite impressive breakthroughs, high hopes and bold promises, the automotive industry has recently acknowledged that “real” self-driving cars aren’t coming anytime soon. In the words of Waymo’s own CEO (Waymo is Google’s self-driving car initiative): “Autonomous cars won’t ever be able to drive in all conditions. And it will be decades before self-driving cars are all over the road.

But just like there are simple marketing activities where “Marketing Automation” excels, there are simple driving conditions where “Driving Automation” excels.

Waymo has launched an experimental robotaxi service in Phoenix, Arizona. Or, more accurately, in a suburb of Phoenix named Chandler. I wasn’t in the room when Waymo picked Chandler as their experimentation ground. But let me guess their line of reasoning. No rain. No hills. No traffic. No pedestrians. And the biggest wild animal is a rattlesnake. Only Elon Musk would have picked Manhattan.

Does your business look like Chandler, Arizona? Or does it look like New York City?

What marketing automation and personalization engines do well is driving in Chandler, Arizona. When you browse for a pair of socks on a retailer’s website, algorithms are thoughtful enough to send you an email reminder. They’re smart enough to show you similar pairs of socks. And stubborn enough to monologue about socks for the coming six months.

But is this really what marketing was supposed to be?

Real Marketing is like driving in Manhattan. You’re never alone on the road. Many stakeholders are interacting with you. Driving conditions are changing all the time. Everybody has a different agenda and you constantly need to readjust your plans. All this requires the kind of situational awareness and adjustability that, for decades to come, will remain the privilege of humans.

To be fair, there are some specialized B2C businesses that actually resemble Chandler, Arizona. If you’re a mobile gaming app or the Dollar Shave Club, your offering is simple enough and stable enough for you to deploy a fully robotized Customer Marketing Machine that covers Sales, Journey and Real Marketing. For you, Marketing Automation may really exist.

But as soon as you have a diverse offering, a diverse customer base or dynamic business needs, it doesn’t really matter whether it truly looks like New York, Paris, Tokyo or Djakarta, or whether it simply looks like Boston, Amsterdam or Sydney. In all cases, it will be a mess. For you, Marketing Automation does not exist.

4. Your biggest problem isn’t a lack of automation

Question 1. Is the majority of your messaging volume generated by campaigns that are led by your business needs and require human input?

“Yes” is the expected answer. There’s nothing wrong with it. It just means that:

  • These campaigns are the de facto core of your customer marketing activity. They are the Real Customer Marketing.
  • They de facto define your brand experience in the eyes of your customers.

It also means that you cannot get rid of these campaigns. If your business could afford it, you would have done it long ago. These campaigns solve real business needs and drive demand. But are you running them well?

Question 2. Is the planning of these campaigns a top-down process, insufficiently informed by customer data?

Most B2C marketers also answer “yes” and they know that this is a problem. All too often, the marketing plan is simply materialized in a spreadsheet that serves as a brief for all production tasks down the line, without any feedback loop or opportunity to adjust the plan based on data.

Some people actually wish that the whole process could just go away. What if a mythical self-driving marketing machine could just crunch all the data? Sure, marketers would lose their jobs, but isn’t it the way of the future?

Well, no. Real Marketing is like driving in Manhattan, not Chandler. It won’t be automated anytime soon. Your key business drivers (sales priorities, product news, partnerships, the notes from the weekly steering meeting) aren’t even accessible in a database. And humans continue to beat machines at storytelling.

Who needs a self-driving car anyway? Not you. All you need is power steering and Waze.

Your biggest problem isn’t a lack of automation. Sure, productivity is an issue. But couldn’t you simply leverage templates? Or improve your CMS? Or fix integration glitches? Or maybe you’re just asking too many marketing questions to your data science team? Your team does need automatic transmission and power steering. But the solutions exist on the market. Productivity issues are common with repetitive, labor-intensive tasks. Most likely, they’re just a symptom of a lack of investment. But it is very rare that the solution to a productivity issue is to get rid of humans altogether.

Your biggest problem is that your Customer Marketing isn’t sufficiently data-driven. A top-down process where product news, promotions and curated content are forced upon millions of consumers based on gut feel and sales priorities cannot be the right way to do marketing in the customer-centric era.

This is where most B2C organizations are lagging and, honestly, it more than justifies the bad reputation of Customer Marketing programs. It also explains why some organizations wrongly assumed that Real Marketing was bound to disappear and cut their investment in marketing operations and production.

Many companies still do batch-and-blast. Others rely on RFM segmentation, which provide useful insights on the health of your database but do not tell you who will be interested by the special offer that you’ll run tomorrow. Propensity models are a better approach. But they view each topic in isolation and offer no guidance on how to prioritize topics and build the master plan.

None of this is up to the task. Data-driven Customer Marketing is several orders of magnitude harder:

  • on the one hand, you have the business inputs (new products, promotions, sales priorities, partnerships)
  • on the other hand, you have millions (or billions) of data points about your customers
  • but there is no direct way to reconcile the two: your mission is to generate demand and therefore you cannot rely on explicit customer intent
  • and it is the entire plan that has to be data-driven, not just isolated campaigns.

The good news is that technology has progressed to a point where this problem can be solved. Deep learning makes it possible to scan millions of unrelated data points and find the hidden structure. This technology powers a new generation of solutions that assist marketers in solving deeper, more strategic problems. You can finally understand the demand patterns across all your customers and your entire offering. You can anticipate a customer’s likelihood to buy a specific item before you even capture any explicit intent from them. You can prioritize based on value, not gut-feel.

From a marketer’s perspective, this is like finally getting the Waze that they need to navigate the complexities of real-life B2C Customer Marketing. They can finally reconcile the dynamic business needs and the conflicting priorities of multiple stakeholders within the organization with the individual preferences and unexpressed demand of a great number of customers. And no worries when your plans suddenly change: the best trajectory is recomputed.

Success stories of Tinyclues clients illustrate what marketers can achieve in real-life conditions with such a solution.

Waze does not replace you. It augments you as a driver. It’s not taking power away from you, it’s giving your superpowers. You’ve become a data-driven driver, but you’re still driving.

Bolder ambitions for Customer Marketing

Sales Automation and Journey Automation are perfect use cases for “Marketing Automation,” but they are the exception. Beyond that, full automation is the wrong paradigm for Customer Marketing.

It is healthy to be frustrated if your newsletter program sucks, but is it reasonable to assume that it will get better once it is written by robots?

Successful AI solutions are defined not just by what they automate but, more importantly, by what they empower you to do better. Automation is the boring part of the AI promise. Augmenting human intelligence is the exciting part.

Taking B2C Customer Marketing to the next level means requires taking it seriously. Personalization isn’t about filling a template with recently viewed items and stuff that ‘you may also like.’ It’s about aligning the deep history of millions of relationships with individual customers, most of whom haven’t interacted with you in months, with what is currently going on with the business and the needs of your organization.

The solution isn’t to create more workflows in your ESP / Marketing Cloud. Or to buy a new ESP / Marketing Cloud. Very likely, you have already implemented most of what “Marketing Automation” can offer. There just isn’t much room for further growth here.

The next opportunity is to improve the decisioning, planning and targeting of your marketing campaigns. Most of these activities currently take place outside of your ESP / Marketing Cloud. This is about informing all decisions with customer data and bridging the intelligence gap.

But the most important change involves resetting the expectations around Customer Marketing.

B2C Customer Marketing isn’t a necessary evil or a remnant of the past.

Customer Marketing is a core strategic function that dialogues with the business, understands the customers, generates new demand and steers it toward where the needs are.

When enlightened by customer data, it massively improves its traditional KPIs. More importantly, it becomes more agile, more reliable and can finally scale. It responds to more demands from the business. And pro-actively engages commerce and category management about untapped marketing opportunities that were detected in the data.

This is the opportunity. With the right ambition and the right technology, Customer Marketing can become a better partner for the entire organization.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade