Marketing in 2019: The Good, Bad, and Ugly

With an increasingly complex data and marketing technology environment, 2019 will be the most challenging year yet for marketing and data teams to cut through the noise and deliver an effective digital experience. With heightened customer expectations, tighter competition, data gaps and a progressively confusing MarTech landscape to navigate, all but the best prepared teams are going to struggle to reach their goals.

While it’s impossible to plan for the unknown — and there are certainly many unknowns in the 2019 macro and micro environment — there are a few trends to pay attention to as we head into the new year. The following is an assessment of these trends and how companies are adapting in the better and worse case scenarios.

What’s happening in the marketing, data and marketing technology ecosystem?

Marketers are shifting focus from targeting audiences to developing a relationship with individual customers at scale. This shift is substantiated by: (1) competitive pressure and resulting customer expectations, (2) proliferation in the availability and use of customer data and (3) regulatory / data privacy concerns.

Competitive Pressure:

Over the past few years, as Facebook and Google have commanded a dominant position in the digital advertising space and the ease of setting up a Shopify business and running Instagram ads has put margin pressure on legacy players, once profitable channels have become increasingly difficult from which to squeeze return. This and elevated consumer expectations on the digital experience a brand should provide have caused brands to shift their strategy from targeting audiences to attempting to piece together data across channels to understand and market to individual customers.

Simultaneously, digital-natives like Amazon, Spotify, Netflix, etc. have built their own first-party data ecosystems which make it nearly impossible for the typical business to compete on the basis of knowing and delivering to customer expectations. These tech juggernauts are years ahead in the shift from audience targeting to known digital customer relationships; they consistently deliver a better digital customer experience than other brands.

Pressure to maintain market share and to also beat competitors in these channels has forced enterprises and growth businesses alike to play catch up in their data strategy. Such businesses have focused on gathering every possible intimate detail on their customers, sometimes at the expense of not using useful data they already have to drive revenue or profit. This has been especially true in larger organizations where consulting firms and enterprise cloud data solutions have preyed upon IT budgets and the desire for an “enterprise-wide single customer view.” Rarely do these initiatives take into account the needs of the team accountable for outcomes.

Proliferation of Customer Data:

As a result of this pressure and the desire to achieve faster time-to-value than a year-plus-long systems migration, we saw the rapid rise of the Customer Data Platform (CDP) space. This space has actually been in existence for more than a decade, but businesses — from tag managers to predictive marketing clouds to MDM platforms to postage technology (actually) — are all taking advantage of the category confusion and market demand and branding themselves as CDPs. This has created a supreme challenge for marketing and technology teams to cut through the noise. RFPs for “CDP vendors” increased ten-fold in 2018 vs. 2017.

Regulatory and Privacy Concerns:

2018 was also the year of GDPR and the announcement that similar legislation would be coming to a dozen U.S. states in the not-too-distant future. It’s clear from these policies that customers are increasingly uncomfortable with the ways businesses use their data to market to them. The bigger challenge for marketing teams than simply complying with the right to be forgotten / erasure, is adopting a more effective owned and cross-channel marketing strategy so that customers feel they have an authentic relationship with the business.

To complicate data privacy matters, a number of businesses encountered high profile data breaches this year. These breaches were lighter fluid on the simmering fire of consumer distrust regarding corporate data use. Remember when I said Facebook is dominating the digital advertising space because of their first-party data assets? Their stock is down over the past twelve months as a result of sketchy use of customer data and ensuing distrust. No time like the present to start selling a connected device that allows you to “video chat” with your “Facebook friends” in your home!

Now, what are companies doing well and not so well when navigating this changing marketing and data landscape?

Multi-channel:

Good: Teams that are recognizing the right channel in which to engage customers in this evolving multi-channel world are winning. In some cases, savvy teams are using engagement metrics from one channel to reach customers in other channels. Do millennials with high LTV/CLV not open your emails? Automate them into an Instagram retargeting sequence and a lookalike model to target similar customers. Do some customers have high owned channel engagement? Suppress some from paid media and measure the impact of not targeting them, and then sub-segment the cohort to see which attributes are associated with needing multi-channel messaging to convert.

Bad: Some customers will never buy no matter how many times you creep them out with Criteo ads. Others may buy often but have low / zero response to marketing in certain channels. Attribution is a tough challenge to solve, but it seems many businesses are still blasting emails, ads, and even messages in more invasive channels (i.e. push / SMS) with limited consistency across channels and limited or no recognition of customer intent.

Data:

Good: Businesses that can tie their data and IT infrastructure needs to use cases and revenue are winning. Some businesses have recognized the low hanging bananas and are able to leverage this data for better targeting and personalization. Examples of this are: the ability to connect retail and e-commerce data as well as clientelling and customer support applications and the ability to marry core first-party customer attributes with behavioral data. This blending of “fast moving” and “slow moving” data allows marketers to personalize an abandon search email, for example, based on a customer’s historical discount sensitivity or to deploy personalized ads after an in-store visit without a purchase.

Bad: Businesses not chasing time-to-value in this competitive market are falling behind. Netflix can show me which shows I might be interested in and Amazon ships me soap before I run out of it, so if you’re a 100-year-old magazine publisher or retailer or a direct-to-consumer flower delivery business, you probably shouldn’t try and build an internal system to aggregate, cleanse, manage, and integrate data across the business. Adam Smith would be disappointed.

AI:

Good: Businesses that privilege data science and bespoke customer-level predictive attributes are winning. Good is being able to understand which customers are likely to churn or repeat a purchase, upgrade their seat or decline in lifetime value. Better is pairing that with marketing responsiveness to understand not only which customers are likely to take an action, but also which customers are likely to respond to a marketing intervention encouraging them to take an action.

Bad: Two things: (1) doing it for the sake of doing it and (2) thinking that it’ll replace effective marketing strategy. It seems many marketers these days are focused on, for example, predicting the gender of a customer in real-time on the website based on their browse behavior and then serving up content tailored to that profile. Seems admirable. The problem? High gender skew in customer profile and prevalence of gifting. Why not just show the customer what they browsed in an email after they visited the website? You don’t need to predict the customer’s gender or know what they’re thinking to remind a customer what they might have wanted to buy. You just need to know what they browsed. Overall, AI is just not at a point where it can replace effective marketing strategy. There is no cloud solution into which you can integrate your data that will spit out money, despite what some vendor at a conference told you.

Personalization:

Good: At the core of good personalization in this increasingly data-driven world is value to the consumer. Customers genuinely enjoy receiving emails with products they might want to buy or reminders when an item they attempted to add to their cart comes back in stock. Businesses that are willing to deploy simple personalization without having to account for every edge case are able to deploy tests much faster, reach higher revenue per customer and reduce opportunity cost. Still worried that your customer might not actually want to know about a co-branded event in their area because their shipping address is different from their billing address and you think they might have a second home? Don’t worry, your customer won’t care and you’ll thank yourself for getting the campaign out the door. We seem to have finally seen the death of “dumb personalization” and the ability to show a snow .gif in email if it’s snowing in a customer’s geo. Thank goodness.

Bad: Personalization can get creepy if there’s not a value exchange or if you’re using probabilistic matching. One way to tell if your personalization strategy is amiss is to follow a simple rule: don’t be weird. Now that businesses that get the data part right have more data than ever at their fingertips, the temptation is to go overboard on personalization. If you’re a cable company, you might know my likelihood to move and where I live and whether I’m likely to enjoy sports. You also definitely know that I’ve called customer support five times in the past three months because I can’t access a premium sports channel I purchased. What’s the better subject line: (a) moving soon? see our sports packages in your area or (B) help with your recent support request? Dear, Verizon Fios: see “Multi-Channel” and “Data.” If you’re targeting ads by tying devices to users with IPs or pixels alone you’ve probably ruined someone’s holiday gifts, surprise vacations, or wedding engagements.

Identity:

Good: Identity has become an interesting space as businesses look to develop a better relationship with their known customers. As a requirement, businesses need to now tie anonymous web visits back to customers and use this information to take action. This can be hugely beneficial in two areas: (1) increasing the volume of owned channel sends to non-authenticated users and (2) enriching the customer profile with otherwise anonymous user data. This is a key space to watch in 2019.

Bad: Borrowing elements from AI and personalization, some businesses buy the buzzword and not the outcome and others use identity probabilistically. There’s nothing wrong with probabilistic matching for paid media or enriching a customer profile. The challenge, as with personalization, is not creeping the customer out or damaging a relationship by getting the match wrong. Some businesses (especially content businesses where there’s a need to login often) aren’t conducting the basic blocking and tackling of identification and force users to re-login every time they visit the site. Not good.

If you’ve hung in this long, you’ve either enjoyed this take or you have views of your own. Either way, I encourage you to get in touch: dylan@simondata.com