Five Principles of Modern Marketing

Every marketer and media company these days is trying to unlock the secret to personalization. Everyone wants to be the next Amazon, anticipating customer wants and desires and delivering real-time customization.

Actually, everyone might need to be an Amazon going forward; Harris Interactive and others tell us that getting customer experience wrong means up to 80% of customers will leave your brand and try another – and it takes seven times more money to reacquire that customer than it did initially.

How important is personalization? In a recent study, 75% marketers of marketers said that there’s no such thing as too much personalization for different audiences, and 94% know that delivering personalized content is important to reaching their audiences.

People want and expect personalization and convenience today, and brands and publishers that cannot deliver it will suffer similar fates. However, beyond advanced technology, what do you need to believe to make this transformation happen? What are the core principles a company needs to adhere to, in order to have a shot at transforming themselves into customer-centric enterprises?

Here are five:

Put People First

It’s a rusty old saw but, like any cliché, it’s fundamentally true. For years, we have taken a very channel-specific view of engagement, thinking in terms of mobile, display, social and video. But those are channels, apps and browsers. Browsers don’t buy anything; people do.

A people-centric viewpoint is critical to being a modern marketer. True people-based marketing needs to extend beyond advertising and start to include things like sales, service and ecommerce interactions – every touchpoint people have with brands.

People – customers and consumers – must reside at the center of everything, and the systems of engagement we use to touch them must be tertiary. This makes the challenges of identity resolution the new basis of competition going forward.

Collect Everything, Measure Everything

A true commitment to personalized marketing means that you have to understand people. For many years, we have assigned outsized importance to small scraps of digital exhaust such as clicks, views and likes as signals of brand engagement and intent. Mostly, they’ve lived in isolation, never informing a holistic view of people and their wants and desires.

Now we can collect more of this data and do so in real time. Modern enterprises need to become more obsessive about valuing data. Every scrap of data becomes a small stitch in a rich tapestry that forms a view of the customer.

We laughed at the “data is the new oil” hyperbole a few years back – simply because nobody had a way to store and extract real value from the sea of digital ephemera. Today is vastly different because we have both the technology and processes to ingest signals at scale – and use artificial intelligence to refine them into gold. Businesses that let valuable data fall to the floor without measuring them might already be dead, but they just don’t know it yet.

Be A Retailer

A lot of brands aren’t as lucky as popular hotel booking sites. To book a room, you need to sign up with your email. Once you become a user, the company collects data on where you like to go, how often you travel, how much you pay for a room and even what kind of mattress you prefer. Any brand would kill for that kind of one-to-one relationship with a customer.

Global CPG brands touch billions of lives every day, yet often have to pay other companies to learn how their marketing spend affected sales efforts. Brands must start to own customer relationships and create one-to-one experiences with buyers. We are seeing the first step with things like Dash buttons and voice ordering, though still through a partner, but we will see this extend even further as brands change their entire business models to start to own the retail relationship with people. The key pivot point will come when brands actually value people data as an asset on their balance sheets.

See The World Dynamically

The ubiquity of data has led to an explosion of microsegmentation. I know marketers and publishers that can define a potential customer to 20 individual attributes. But people can go from a “Long Island soccer mom” on Monday to an “EDM music lover” on Friday night. Today’s segmentation is very much static – and very ineffective for a dynamic world where things change all the time.

To get the “right message, right place, right time” dynamic right, we need to understand things like location, weather, time of day and context – and make those dynamic signals part of how we segment audiences. To be successful, marketers and media companies must commit to thinking of customers as the dynamic and vibrant people they are and enable the ability to collect and activate real-time data into their segmentation models.

Think Like A Technologist

Finally, to create the change described above requires a commitment to understanding technology. You can’t do “people data” without truly understanding data management technology. You can’t measure everything without technology that can parse every signal. To be a retailer, you have to give customers a reason to buy directly from you. Thinking about customers dynamically requires real-time systems of collection and activation.

But technology and the people to run it are expensive investments, often taking months and years to show ROI, and the technology changes at the velocity of Moore’s Law. It’s a big commitment to change from diaper manufacturer to marketing technologist, but we are starting to understand that it is the change required to survive an era where people are in control.

Some say that it wasn’t streaming media technology that killed Blockbuster, but the fact that people hated their onerous late fees. It was probably both of those things. Tomorrow’s Blockbusters will be the companies that cannot apply these principles of modern, personalized marketing – or do not want to make the large investments to do so.

[This article originally appeared in AdExchanger on 8/7/2017.]

Advertisements

DMPs are Dead. Long Live DMPs.

 

King

Much like latter day King of Rock and Roll, Elvis Presley, today’s ubiquitous data management platforms will eventually die as an independent buying category but live on in the greater consciousness. And karate. 

Gartner’s Marty Kihn recently made an argument that ad tech and mar tech would not come together, contrary to what he had predicted a few years ago. When Marty speaks about ad tech, people listen.

 

Like many people, when I read the headline, I thought to myself, “That makes no sense!” But those who read the article more closely understand that the disciplines of ad tech and mar tech will certainly be bound closer together as systems align – but the business models are totally incompatible.

Advertising technology and the ecosystem that supports it, both from a commercial business model perspective (percentage of media spend billed in arrears) and the strong influence of agencies in the execution process, has meant that the alignment with software-as-a-service (SaaS) marketing technology is not just an engineering problem to solve.

Marketing leaders and brands need to change the way they do their P&L and budgeting and reevaluate business process flows both internally and with outside entities such as agencies to ensure that even if the technology may be right, the execution needs to be optimal to achieve the desired results.

There are also plenty of technical hurdles to overcome to truly integrate mar tech and ad tech – most notably, finding a way to let personally identifiable information and anonymous data flow from system to system securely. While those technical problems may be overcome through great software engineering, the business model challenge is a more significant hurdle.

I remember getting some advice from AdExchanger contributor Eric Picard when we worked together some years ago. I was working at a company that had a booming ad tech business with lots of customers and a great run rate, operating on the typical ad network/agency percentage-of-spend model.

At the time, we were facing competition from every angle and getting disrupted quickly. Eric’s suggestion was to transform the company to a platform business, license our technology for a fixed monthly fee and begin to build more predictable revenues and a dedicated customer base. That would have meant parting ways with our customers who would not want to pay us licensing fees and rebuilding the business from scratch.

Not an easy decision, but one we should have taken at the time. Eric was 100% right, but transforming a “run rate” revenue ad tech business into a SaaS business takes a lot of guts, and most investors and management didn’t sign up for that in the first place.

This is a long way of saying that Marty is right. There are tons of ad tech businesses that simply cannot transform themselves into marketing software stacks, simply because it requires complete change – from a structural financial perspective (different business model) and a people perspective (different sales skills required).

[This post appeared in AdExchanger on 5/9/2017]

What is the future of DMPs?

In the 1989 film “Back to the Future II,” Marty McFly traveled to Oct. 21, 2015, a future with flying cars, auto-drying clothes and shoes that lace automatically.

What is the future of data management platforms? This is a question I get asked a lot.

The short answer is that DMPs are now part of larger marketing stacks, and brands realize that harnessing their data is a top priority in order to deliver more efficient marketing.

This is a fast-moving trend in which companies are licensing large enterprise stacks and using systems integrators to manage all marketing—not just online advertising.

As detailed in Ad Age (Marketing clouds loom), the days of turning to an agency trade desk or demand side platform (DSP) to manage the “digital” portions of advertising are fading rapidly as marketers are intent on having technology that covers more than just advertising.

Building consumer data platforms

A few years ago, a good “stack” might have been a connected DMP, DSP and ad server. A really good stack would feature a viewability vendor and start a dynamic creative optimization (DCO). The focus then was on optimizing for the world of programmatic buying and getting the most out of digital advertising as consumers’ attention shifted online, to mobile and social, rather than television.

Fast forward a few years, and the conversations we are having with marketers are vastly different. As reported in AdExchanger, more than 40% of enterprise marketers license a DMP, and another 20% will do so within the next 12 months. DMP owners and those in the market for one are increasingly talking about more than just optimizing digital ads. They want to know how to put email marketing, customer service and commerce data inside their systems. They also want data to flow from their systems to their own data lakes.

Many are undertaking the process of building internal consumer data platforms (CDPs), which can house all of their first-party data assets—both known and pseudonymous user data.

We are moving beyond ad tech. Quickly.

Today, when those in the market are considering licensing a “DMP” they are often thinking about “data management” more broadly. Yes, they need a DMP for its identity infrastructure, ability to connect to dozens of different execution systems and its analytical capabilities. But they also need a DMP to align with the systems they use to manage their CRM data, email data, commerce systems, and marketing automation tools.

Data-driven marketing no longer lives in isolation. After I acquire a “luxury sedan intender” online, I want to retarget her—but I also want to show her a red sedan on my website, e-mail her an offer to come to the dealership, serve her an SMS message when she gets within range of the dealership to give her a test drive incentive, and capture her e-mail address when she signs up to talk to a salesperson. All of that needs to work together.

Personalization demands adtech and martech come together

We live in a world that demands Netflix and Amazon-like instant gratification at all times. It’s nearly inconceivable to a Millennial or Generation Z if a brand somehow forgets that they are a loyal customer because they have so many choices and different brands that they can switch to when they have a bad experience.

This is a world that requires adtech and martech to come together to provide personalized experiences—not simply to create more advertising lift, but as the price of admission for customer loyalty.

So, when I am asked, what is the future of DMPs, I say that the idea of licensing something called a “DMP” will not exist in a few years.

DMPs will be completely integrated into larger stacks that offer a layer of data management (for both known and unknown data) for the “right person;” an orchestration layer of connected execution systems that seek to answer the “right message, right time” quandary; and an artificial intelligence layer, which is the brains of the operation trying to figure out how to stitch billions of individual data points together to put it all together in real time.

DMPs will never be the same, but only in the sense that they are so important that tomorrow’s enterprise marketing stacks cannot survive without integrating them completely, and deeply.

[This post was originally published 11 May, 2017 by Chris O’Hara in Econsultancy blog]

The Technology Layer Cake

spumoni-layer-cakeI saw a great presentation at this year’s Industry Preview where Brian Anderson of LUMA Partners presented on the future of marketing clouds. His unifying marketechture drawings looked like an amalgamation of various whiteboarding sessions I have had recently with big enterprise marketers, many of whom are building the components of their marketing “stacks.” Marketers are feverishly licensing offerings from all kinds of big software companies and smaller adtech and martech players to build a vision that can be summed up like this:

The Data Management Layer

Today’s “stack” really consists of three individual layers when you break it down. The first layer, Data Management (DM), contains all of the “pipes” used to connect people identity together. Every cloud needs to take data in from all kinds of sources, such as internet cookies, mobile IDs, hashed e-mail identity keys, purchase data, and the like. Every signal we can collect results in a richer understanding of the customer, and the DM layer needs access to rich sets of first, second, and third-party data to paint the clearest picture.

The DM layer also needs to tie every single ID and attribute collected to an individual, so all the signals collected can be leveraged to understand their wants and desires. This identity infrastructure is critical for the enterprise; knowing that you are the same guy who saw the display ad for the family minivan, and visited the “March Madness Deals” page on the mobile app goes a long way to attribution. But the DM layer cannot be constrained by anonymous data. Today’s marketing stacks must leverage DMPs to understand pseudonymous identity, but must find trusted ways to mix PII-based data from e-mail and CRM systems. This latter notion has created a new category—the “Customer Data Platform” (CDP), and also resulted in the rush to build data lakes as a method of collecting a variety of differentiated data for analytics purposes.

Finally, the DM layer must be able to seamlessly connect the data out to all kinds of activation channels, whether they are e-mail, programmatic, social, mobile, OTT, or IOT-based. Just as people have many different ID keys, people have different IDs inside of Google, Facebook, Pinterest, and the Wall Street Journal. Connecting those partner IDs to an enterprises’ universal ID solves problems with frequency management, attribution, and offers the ability to sequence messages across various addressable channels.

You can’t have a marketing cloud without data management. This layer is the “who” of the marketing cloud—who are these people and what are they like?

The Orchestration Layer

The next thing marketers need to have (and they often build it first, in pieces) is an orchestration layer. This is the “When, Where, and How” of the stack. E-mail systems can determine when to send that critical e-mail; marketing automation software can decide whether to put someone in a “nurture” campaign, or have a salesperson call them right away; DSPs decide when to bid on a likely internet surfer, and social management platforms can tell us when to Tweet or Snap. Content management systems and site-side personalization vendors orchestrate the perfect content experience on a web page, and dynamic creative optimization systems have gotten pretty good at guessing which ad will perform better for certain segments (show the women the high-heeled shoe ad, please).

The “when” layer is critical for building smart customer journeys. If you get enough systems connected, you start to realize the potential for executing on the “right person, right message, right time” dynamic that has been promised for many years, but never quite delivered at scale. Adtech has been busy nailing the orchestration of display and mobile messages, and the big social platforms have been leveraging their rich people data to deliver relevant messages. However, with lots of marketing money and attention still focused on e-mail and broadcast, there is plenty of work to be done before marketers can build journeys that feature every touchpoint their customers are exposed to.

Marketers today are busy building connectors to their various systems and getting them to talk to each other to figure out the “when, where, and how” of marketing.

The Artificial Intelligence Layer

When every single marketer and big media company owns a DMP,and has figured out how to string their various orchestration platforms together, it is clear that the key point of differentiation will reside in the AI layer. Artificial intelligence represents the “why” problem in marketing—why am I e-mailing this person instead of calling her? Should I be targeting this segment at all? Why does this guy score highly for a new car purchase, and this other guy who looks similar doesn’t? What is the lifetime value of this new business traveler I just acquired?

While the stacks have tons of identity data, advertising data, and sales data, they need a brain to analyze all of that data and decide how to use it most effectively. As marketing systems become more real-time and more connected to on-the-go customers than ever before, artificial intelligence must drive millions of decisions quickly, gleaned from billions of individual data points. How does the soda company know when to deliver an ad for water instead of diet soda? It requires understanding location, the weather, the person, and what they are doing in the moment. AI systems are rapidly building their machine learning capabilities and connecting into orchestration systems to help with decisioning.

All Together Now

The layer cake is a convenient way to look at what is happening today. The vision for tomorrow is to squish the layer cake together in such a way that enterprises get all of that functionality in a single cake. In four or five years, every marketing orchestration system will have some kind of built-in DMP—or seamless connections to any number of them. We see this today with large DSPs; they all need an internal data management system for segmentation. Tomorrow’s orchestration systems will all have built-in artificial intelligence as a means for differentiation. Look at e-mail orchestration today. It is not sold on its ability to deliver messages to inboxes, but rather on its ability to provide that service in a smarter package to increase open rates and provide richer analytics.

It will be fun to watch as these individual components come together to form the marketing clouds of the future. It’s a great time to be a data-driven marketer!

[This post was originally published April 4, 2017 on Econsultancy blog

Deepening The Data Lake: How Second-Party Data Increases AI For Enterprises

chrisohara_managingdata_updated

I have been hearing a lot about data lakes lately. Progressive marketers and some large enterprise publishers have been breaking out of traditional data warehouses, mostly used to store structured data, and investing in infrastructure so they can store tons of their first-party data and query it for analytics purposes.

“A data lake is a storage repository that holds a vast amount of raw data in its native format until it is needed,” according to Amazon Web Services. “While a hierarchical data warehouse stores data in files or folders, a data lake uses a flat architecture to store data.”

A few years ago, data lakes were thought to be limited to Hadoop applications (object storage), but the term is now more broadly applied to an environment in which an enterprise can store both structured and unstructured data and have it organized for fast query processing. In the ad tech and mar tech world, this is almost universally about first-party data. For example, a big airline might want to store transactional data from ecommerce alongside beacon pings to understand how often online ticket buyers in its loyalty program use a certain airport lounge.

However, as we discussed earlier this year, there are many marketers with surprisingly sparse data, like the food marketer who does not get many website visitors or authenticated customers downloading coupons. Today, those marketers face a situation where they want to use data science to do user scoring and modeling but, because they only have enough of their own data to fill a shallow lake, they have trouble justifying the costs of scaling the approach in a way that moves the sales needle.

chris_ohara1

Figure 1: Marketers with sparse data often do not have enough raw data to create measureable outcomes in audience targeting through modeling. Source: Chris O’Hara.

In the example above, we can think of the marketer’s first-party data – media exposure data, email marketing data, website analytics data, etc. – being the water that fills a data lake. That data is pumped into a data management platform (pictured here as a hydroelectric dam), pumped like electricity through ad tech pipes (demand-side platforms, supply-side platforms and ad servers) and finally delivered to places where it is activated (in the town, where people live).

As becomes apparent, this infrastructure can exist with even a tiny bit of water but, at the end of the cycle, not enough electricity will be generated to create decent outcomes and sustain a data-driven approach to marketing. This is a long way of saying that the data itself, both in quality and quantity, is needed in ever-larger amounts to create the potential for better targeting and analytics.

Most marketers today – even those with lots of data – find themselves overly reliant on third-party data to fill in these gaps. However, even if they have the rights to model it in their own environment, there are loads of restrictions on using it for targeting. It is also highly commoditized and can be of questionable provenance. (Is my Ferrari-browsing son really an “auto intender”?) While third-party data can be highly valuable, it would be akin to adding sediment to a data lake, creating murky visibility when trying to peer into the bottom for deep insights.

So, how can marketers fill data lakes with large amounts of high-quality data that can be used for modeling? I am starting to see the emergence of peer-to-peer data-sharing agreements that help marketers fill their lakes, deepen their ability to leverage data science and add layers of artificial intelligence through machine learning to their stacks.

chris_ohara2

Figure 2: Second-party data is simply someone else’s first-party data. When relevant data is added to a data lake, the result is a more robust environment for deeper data-led insights for both targeting and analytics. Source: Chris O’Hara.

In the above example (Figure 2), second-party data deepens the marketer’s data lake, powering the DMP with more rich data that can be used for modeling, activation and analytics. Imagine a huge beer company that was launching a country music promotion for its flagship brand. As a CPG company with relatively sparse amounts of first-party data, the traditional approach would be to seek out music fans of a certain location and demographic through third-party sources and apply those third-party segments to a programmatic campaign.

But what if the beer manufacturer teamed up with a big online ticket seller and arranged a data subscription for “all viewers or buyers of a Garth Brooks ticket in the last 180 days”? Those are exactly the people I would want to target, and they are unavailable anywhere in the third-party data ecosystem.

The data is also of extremely high provenance, and I would also be able to use that data in my own environment, where I could model it against my first-party data, such as site visitors or mobile IDs I gathered when I sponsored free Wi-Fi at the last Country Music Awards. The ability to gather and license those specific data sets and use them for modeling in a data lake is going to create massive outcomes in my addressable campaigns and give me an edge I cannot get using traditional ad network approaches with third-party segments.

Moreover, the flexibility around data capture enables marketers to use highly disparate data sets, combine and normalize them with metadata – and not have to worry about mapping them to a predefined schema. The associative work happens after the query takes place. That means I don’t need a predefined schema in place for that data to become valuable – a way of saying that the inherent observational bias in traditional approaches (“country music fans love mainstream beer, so I’d better capture that”) never hinders the ability to activate against unforeseen insights.

Large, sophisticated marketers and publishers are just starting to get their lakes built and begin gathering the data assets to deepen them, so we will likely see a great many examples of this approach over the coming months.

It’s a great time to be a data-driven marketer.

Follow Chris O’Hara (@chrisohara) and AdExchanger (@adexchanger) on Twitter.

How AI will Change UX

user_experience_sarah_weise In 1960, the US Navy coined a design principle: Keep it simple, stupid.

When it comes to advertising and marketing technology, we haven’t enjoyed a lot of “simple” over the last dozen years or so. In an increasingly data-driven world where delivering a relevant customer experience makes all the difference, we have embraced complexity over simplicity, dealing in acronyms, algorithms and now machine learning and artificial intelligence (AI).

When the numbers are reconciled and the demand side pays the supply side, what we have been mostly doing is pushing a lot of data into digital advertising channels and munching around the edges of performance, trying to optimize sub-1% click-through rates.

That minimal uptick in performance has come at the price of some astounding complexity: ad exchanges, third-party data, second-price auctions and even the befuddling technology known as header bidding. Smart, technical people struggle with these concepts, but we have embraced them as the secret handshake in a club that pays it dues by promising to manage that complexity away.

Marketers, however, are not stupid. They have steadily been taking ownership of their first-party data and starting to build marketing tech stacks that attempt to add transparency and efficiency to their outbound marketing, while eliminating many of the opaque ad tech taxes levied by confusing and ever-growing layers of licensed technology. Data management platforms, at the heart of this effort to take back control, have seen increased penetration among large marketers – and this trend will not stop.

This is a great thing, but we should remember that we are in the third inning of a game that will certainly go into extra innings. I remember what it was like to save a document in WordPerfect, send an email using Lotus Notes and program my VCR. Before point-and-click interfaces, such tasks were needlessly complex. Ever try to program the hotel’s alarm clock just in case your iPhone battery runs out? In a world of delightful user experience and clean, simple graphical interfaces, such a task becomes complex to the point of failure.

Why Have We Designed Such Complexity Into Marketing Technology?

We are, in effect, giving users who want big buttons and levers the equivalent graphical user interface of an Airbus A380: tons of granular and specific controls that may take a minute to learn, but a lifetime to master.

How can we change this? The good news is that change has already arrived, in the form of machine learning and artificial intelligence. When you go on Amazon or Netflix, do you have to program any of your preferences before getting really amazing product and movie recommendations? Of course not. Such algorithmic work happens on the back end where historical purchases and search data are mapped against each other, yielding seemingly magical recommendations.

Yet, when airline marketers go into their ad tech platform, we somehow expect them to inform the system of myriad attributes which comprise someone with “vacation travel intent” and find those potential customers across multiple channels. Companies like Expedia tell us just what to pay for a hotel room with minimal input, but we expect marketers to have internal data science teams to build propensity models so that user scores can be matched to a real-time bidding strategy.

One of the biggest trends we will see over the next several years is what could be thought of as the democratization of data science. As data-driven marketing becomes the norm, the winners and losers will be sorted out by their ability to build robust first-party data assets and leverage data science to sift the proverbial wheat from the chaff.

This capability will go hand-in-hand with an ability to map all kinds of distinct signals – mobile phones, tablets, browsers, connected devices and beacons – to an actual person. This is important for marketers because browsers and devices never buy anything, but customers do. Leading-edge companies will depend on data science to learn more about increasingly hard-to-find customers, understand their habits, gain unique insights about what prompts them to buy and leverage those insights to find them in the very moment they are going to buy.

In today’s world, that starts with data management and ends with finding people on connected devices. The problem is that executing is quite difficult to automate and scale. Systems still require experts that understand data strategy, specific use cases and the value of an organization’s siloed data when stitched together. Plus, you need great internal resources and a smart agency capable of execution once that strategy is actually in place.

However, the basic data problems we face today are not actually that complicated. Thomas Bayes worked them out more than 300 years ago with a series of probabilistic equations we still depend on today. The real trick involves packaging that Bayesian magic in such a way that the everyday marketer can go into a system containing “Hawaiian vacation travel intenders” for a winter travel campaign and push a button that says, “Find me more of these – now!”

Today’s problem is that we depend on either a small amount of “power users” – or the companies themselves – to put all of this amazing technology to work, rather than simply serving up the answers and offering a big red button to push.

A Simpler Future For Marketers?

Instead of building high-propensity segments and waiting for users to target them, tomorrow’s platforms will offer preselected lists of segments to target. Instead of having an agency’s media guru perform a marketing-mix model to determine channel mix, mar tech stacks will simply automatically allocate expenditures across channels based on the people data available. Instead of setting complex bid parameters by segment, artificial intelligence layers will automatically control pricing based on bid density, frequency of exposure and propensity to buy – while automatically suppressing users who have converted from receiving that damn shoe ad again.

This is all happening today, and it is happening right on time. In a world with only tens of thousands of data scientists and enough jobs for millions of them, history will be written by the companies clever enough to hide the math on the server side and give users the elegance of a simple interface where higher-level business decisions will be made.

We are entering into a unique epoch in our industry, one in which the math still rules, but the ability of designers to make it accessible to the English majors who run media will rule supreme.

It’s a great time to be a data-driven marketer! Happy New Year.

Follow Chris O’Hara (@chrisohara) and AdExchanger (@adexchanger) on Twitter.