2024 Predictions are in… What Now?  (Part One - Get Honest About Your Data)

Marketing is going through a generational reset sparked by consumer privacy, digital id signal loss, e-commerce, retail media, clean room proliferation, AI potential, and more. How do you go about ensuring your enterprise marketing data is ready for this new era?  Where do you start?  Where do you end?  How do you make this assessment manageable? Understandable? Actionable? Investable? Well, the very first decision to make is deciding to do this. Really doing it. And then just start!

Somewhat reassuring is the fact that despite the massive increase in marketing fragmentation and complexity, the core elements have not changed that much.  You still have products that you sell to customers at ‘stores’ for a given price and invest in campaigns to deliver messages or incentives via media to drive sales and profit. Simple, right? I suggest that you start there and organize what may seem like chaos into those core components regardless of how disconnected they currently are. This will allow hidden synergies, opportunities, and alternatives to more easily surface later.  

Yes, this seems hard, but it will only get harder as data gets further carved up across media and retailer clean rooms. And perhaps more disruptive will be that media campaigns themselves will also get carved up into multiple audience targeting and activation techniques to achieve the scale and performance required. That said, the time is now.  We have reached our tipping point. Inaction is not an option. The can kickers will not survive. 

The below is a loose step by step approach you can adapt to get organized about what you have, do not have, may soon not have, should have, need to share, need to protect, need to fix, etc.  I am sure there are many ways to tackle this, but I have successfully used variations of this approach before.

1) Inventory your marketing data – Catalog all the marketing data you use, regardless of whether you own it or license it or where it physically resides. Start by understanding all the data you leverage about your customers (name, location, emails, IP addresses, segment assignments, demographics, etc.) and their behaviors (purchasing, media consumption, ad exposures, etc.).  Ditto for data on your products, retailers, campaigns, media platforms, geographies, etc.  For each data type capture key identifiers, descriptors, human assigned attributes, analytically curated attributes, hierarchies, time series behaviors, etc. Pay particular attention to the identifiers used for each data type (hashed email, cookie, IP address, alternative id, store number, DMA, date, etc) as they control your ability to join, link, and map your data across your marketing enterprise and partner platforms.

Note that this activity is not meant to replace the detailed analysis that your data governance and modeling teams do. This data inventory should be reasonably high-level as this is a strategic exercise, not a data architecture exercise.

2) Inventory your marketing use cases - Similar to the above, now catalog all your current marketing strategy, design, execution, and measurement tactics. These use cases should span everything from understanding what your customers buy and why to all the omni-channel marketing tactics that you use to influence their behavior.  Be sure to break down each media campaign into its component parts, like objective setting, message development, target audience creation, campaign setup, bid optimization, performance measurement, etc.  Some of you may be thinking, ‘whoa, that is overkill for this exercise’. Perhaps it is, but then consider what the ideal data requirements are for each stage of a media channel campaign.  They each have unique demands. For example, the data scale, quality, and latency requirements for determining your CTV brand target audience are radically different from the data required to optimize your in-flight CTV bidding strategy.  Breaking these marketing actions into their component parts will help you uncover cross-media synergies and opportunities in the steps below.

3) Connect the marketing data and use case dots - Create a matrix of what specific data is used for each specific marketing use case.  Yes, this will be a very big matrix, especially for marketing organizations who have decentralized their media platforms or companies who have under-invested in their data strategy and architecture.  Stick with it. This step is critical to surfacing and socializing where your best data, best insights, and best practices are under-leveraged or artificially sequestered across your marketing enterprise.

Also, as you do this, you will surely uncover that you still rely on human expertise or even gut in some marketing use cases. Gasp! This is especially important information to collect as it is often unknown and will be the key dependency for your AI driven efficiencies. I suggest you treat these human inputs as data or ‘expert data’.

4) Identify marketing use case risks Up until now we were just collecting baseline facts. Now is the time to thoughtfully review each use case’s industry threats, data dependencies, and associated risks.  A simple red-yellow-green with commentary will do.  I suggest you assess risk in three categories, although you may have more.  The key risk categories being scale, quality, and economic

By scale risk, we are talking about things like effective reach and relevant sample size.  One obvious industry trigger for this is the loss of addressable digital ids via third-party cookies, IP addresses, and mobile ad ids.  Other scale loss could come from more wall gardens, legislation, etc.

By quality risk, you need to make an informed assessment of your data’s true accuracy and representativity.  Is the audience really who you say it is? Are the digital identities current? Was the ad even seen? How many times have they seen the ad already? Was the impression fraudulent? Was the consumer really a non-buyer? 

By economic risk, gauge the likelihood that specific media inventory costs will increase relative to other viable options. Will the increased scarcity of opted-in digital ids or trustworthy deterministic data drive up costs and destroy your ROAS? Would less accurate, but much cheaper ‘best guess’ targeting make more sense? 

5) Add missing marketing data and use cases - This is where you add new things you know you need, but have been putting off.  This includes things like testing Google’s privacy sandbox, contextual targeting, real-time bid price optimization, Gen AI content creation, cross media clean room measurement, etc.  This is also where you add the new ideas and gaps sparked by going through this exercise.  If nothing new came to mind, congrats, but more likely, you did not take this self-reflection seriously or deeply enough.

Once you add the new marketing data sources and use cases, repeat step four and add risk assessments for those too. 

6) Identify mandates, opportunities, and synergies – Now is the time to imagine all the possibilities and dream unrealistic dreams.  This effort will help you establish your marketing data strategy ‘North Star’.  Of course, you pretty much have no chance to get funding for that vision in one go, but by having your enterprise vision thought through and socialized, you will find that opportunities will more easily emerge that will enable you to chip away at the greater good.  Without a visible and aligned North Star vision, each initiative’s own siloed ‘minimally viable product' will pass right over chances to streamline, unify, and benefit the broader marketing enterprise.   

Each of the major 2024 imperatives are ripe with cross-marketing synergies.  For example, Google’s third-party cookie deprecation plan and TOPICS API require understanding each brand’s affinity with their media content taxonomy.  Rather than doing this in a silo, consider partnering with the contextual targeting needs for TV, CTV, OOH, Search, Social, etc. Each has interest, hobby, cause, genre, lifecycle, and moment learnings to share leveraging their own media consumption, survey, and purchasing data. Remember, we are all talking about the same customer!

Another very real example for 2024 is the pursuit for cost efficiencies using Gen AI and other sophisticated machine learning. The tough reality on this however is that most of the expected cost efficiencies and automations come from the elimination of expensive human involvement.  The competing business reality is that the cost of mistakes is very high on many marketing use cases. AI driven platforms will need to inject rules, norms, and guardrails from those same expensive humans (i.e. ‘expert data’). A happy medium needs to be created and rules, norms, and guardrails will need to be converted into ‘data’ and that same data will surely be useful across the enterprise.   Ditto for AI’s greater need for data cleaning, classifying, feature engineering, etc.  Pay it forward.

7) Prioritize your needs and desires – This is the easiest step to say, but perhaps the toughest to do and stick with.  Securing data investments will be difficult in this environment, but it will ultimately be necessary. Doing this analysis should help you credibly prioritize and package data investments as core enablers of your corporate objectives and industry imperatives.  Being credible however, also means lowering priorities and investment for declining or unproven marketing use cases. Be sure to prioritize data investments for individual use cases and for the enterprise as a whole. Data with modest financial impact on an individual use case, but found in many other enterprise use cases, should given full aggregate ROI credit.

8) Summarize your findings and recommendations – This step is too unique to each company and situation for me to recommend much.  The key thing is to present a plan that business folks can understand, rally around, and actually say ‘yes’ to.  Keep the risks and opportunities you are addressing business centric, quantifying wherever possible.   Also, as you proceed with your improved data strategy, be sure to collect and share your successes along the way, especially those that span marketing silos. You want to sell the value of enterprise data collaboration and joint investment.

9) Reflect on this assessment process - How difficult was it to complete? If it took way more people or time than expected, that alone is telling.  If the implications to the consumer privacy, AI capabilities, or retailer media were a surprise to many, that too is telling.  It tells you that you must educate and socialize your new reality better across your company.  Partner with your data governance team who have similar passions on this topic.  Inject some external expertise who have done this before, know industry best practices, and have outside industry credibility to help you sell your vision.

There you go.  Hope it was helpful. Stay tuned for part two of this blog post where I share what new and old data sources will be critical in this new era.

Previous
Previous

2024 Predictions are in… What Now? (Part Two – A New Era Spawns New Data Needs)

Next
Next

More thoughts on their way