As marketing technology, or martech, becomes more sophisticated, companies are finding they can accelerate and optimize everything from basic commercial functions like e-commerce conversion to more complex ones, such as managing multi-dimensional customer relationships across multiple channels. In this interview, Scott Brinker, VP, Platform Ecosystems, at HubSpot, and Jason Heller, a McKinsey partner leading our Digital Marketing Operations and Technology service line, share their insights on the evolving world of martech with McKinsey’s Barr Seitz.
The most important building blocks of the marketing operating model
Scott Brinker: When building a standardized and coherent operating model that can operate across systems, you need to view it through two lenses: technology and data. Data integration is one of the most crucial components. But once you have the pipes connected, you need to synchronize activity and delivery of the insights from that data across many channels to many stakeholders at an accelerated rate.
That said, governance is the most important thing, and often the least-developed piece. You must establish a model of rights that specifies who is able to have access and make changes. You also need to figure out how to manage the process of consistently updating the marketing program framework. This sort of activity used to occur on an ad hoc basis or was kept at arm’s length with service providers, but it now requires a governance plan.
Jason Heller: I see that too. Governance is one of the most important and least-developed pieces of the puzzle, but it’s also important to know what the puzzle is. For me, the marketing operating model is fundamentally made up of three pieces: data and tech enablement, execution via agile marketing, and talent. Today, with the rise of customer-data platforms, using big data and machine learning in the cloud makes it so much easier to unify data so you can create microsegments and manage customer signals that can drive experiences and create value across channels. That core set of unified data sits at the core of the modern marketing capability, but it still requires an agile execution capability and the right people working together to make it all happen.
Meet your new MOM (Marketing Operating Model)
How to drive new growth quickly with martech
Jason Heller: Typically, we find clients tend to be technology rich but insights poor. They may have a data-management platform (DMP), but they aren’t connecting data properly or using the DMP across channels to execute against the highest-value use cases. They have an A/B testing platform but aren’t running rapid iterative tests. They have analytics in place but aren’t looking at the granular insights that spark a series of hypothesis-driven experiments. Companies are not only underutilizing their martech capabilities, they’re also making too many decisions based on tribal knowledge or gut instinct rather than on data. We find companies can remedy this by deploying a process of discovery analytics to identify the “leaky buckets”—where in the journey customers abandon the process and which parts of the journey don’t live up to their potential to create value. In particular, the companies that regularly map granular customer journeys and understand where changes in engagement or conversion rate can make a massive difference tend to mobilize to capture this value. This is also the fastest path to value. Generally, it helps to know where to point your biggest guns.
Scott Brinker: I agree that most companies only use a tiny fraction of their tools. That’s partly because marketing leaders have the least time to understand the tools to see what experiments can be run. The ability to match insights to existing tool capabilities exists further down the organization, which is why you need to implement agile processes and ways of working so people at a lower level can experiment within a framework without creating chaos. You need to allow them to push forward and work with these tools to understand what they’re capable of. That’s when you can find new value pretty quickly. Martech’s value in creating new products, services, and business models
Scott Brinker: Martech has created a revolution, with marketing going from the business of communications to the business of experiences, from bespoke content to bespoke software. You’re taking software and trying to create services through channels to customers.
Some people make fun of growth hacking, but growth hackers build their marketing mechanisms into the product to pull customers through a journey from trial to premium, or leverage their social networks to pull people in. And while not every site is social, the principle of building marketing mechanisms into the product itself is an exciting opportunity.
Jason Heller: Yes, and what’s more, martech has enabled marketers to engage customers directly and at a depth that’s new to most companies. This means that every company has the ability to learn more about its customers by engaging them, which in turn helps shape the products and services brought to market for specific segments. Increasingly, martech platforms are incorporating AI—machine learning, natural-language processing, and computer vision—to derive customer interests in order to drive personalization.
Martech platforms are central to creating and delivering differentiated and relevant experiences by activating customer data at scale. This capability fuels a customer-acquisition engine for a new product or service, it can drive new types of subscriptions, and it helps to manage the upsell, cross-sell, and retention efforts throughout the customer’s lifecycle.
New world of speed and agility
Jason Heller: The automation of marketing activities enables people to operate in agile ways (that is, by automatically analyzing customer responses to new product features) and frees people up to focus on a bigger and more strategic framework of solutions.
Tech can do more and do it faster, but tech can’t do it alone. You need talent, agile practices, and systems that are creating the right feedback loops so that marketing experiments can produce statistically valid feedback in a day or a week. You can then launch dozens or hundreds of tests in the course of a month. Not all the tests will be winners, but with this throughput, the winners can represent hundreds of millions or billions in revenue for a large company.
Scott Brinker: It’s hard for humans to wrap their heads around exponential changes in speed. Things that used to take us days or months to change can now change in seconds. What’s incredible is the relentless parallel automation that computing gives us. For example, an AI program mastered chess by playing billions of games in just four hours. That’s amazing. Similarly, automated optimization programs can drive a tremendous amount of insights and value in a very short time.