Data Activation is Hard: Is it a People, Process, or System issue?

In a recent study, sponsored by SAS, MIT Sloan Management Review surveyed 1,919 marketing leaders around the world, only 49% of respondents reported being able to use data to guide marketing strategy.

The question is WHY?

1. TECHNOLOGY: WHAT WE DO STUFF WITH - Is there a missing platform that could solve these problems? e.g. a CDP- customer data platform. (I'm assuming in this article you have MA+CRM+BI, those are must-have)
2. PEOPLE: WHO'S DOING STUFF - Or is it that the existing technology are not being maximized because the talent on the team doesn't have the know-how? Perhaps they don't have the technical expertise, as well.
3. PROCESS: HOW STUFF IS DONE - Or is it that the organization isn't aligned on the value marketing can play and or the important components of reaching success, like ongoing testing? Perhaps leadership doesn't allow marketing to go here?

I believe it's #2 and #3. Not #1.

I agree Data to Activation is not easy... Here are some thoughts to advance this in an organization.

So first to address - why are so many marketers not doing this? It's NOT because they have the wrong platforms. This CAN be done with MA + CRM + BI, but we need to address #2 and #3 above....

If it's a #3 PROCESS issue - that's a bigger dialogue with C-Suite on the value marketing can add towards growth and a major mindset shift.

If it's a #2 PEOPLE issue - here are components to consider to advance your data activation.

1. START with dialogue on who are your best prospects? Ask sales - who are the best leads and who are the worst leads? See if there are patterns. Look at the data (See #2). This is both a qualitative and quantitative exercise. See if additional data - for example, what technology platform the company is using, would be beneficial... Identify what TRIGGERS occur right before they are ready to lean in and consider new options?
2. Consider predictive modeling - Based on historical data, who are your best prospects? You can define this as highest response to becoming a lead (this is the marketing model), highest value (what do the customers look like that spend the most?), or the most common: highest likelihood to CLOSE and become a customer (this is based on the lead universe (so keep that in mind, that there may be bias there) - what do the records look like that turn into customers?). Which modeling exercise you do depends on your goals and where in the funnel you are optimizing.. What's important here is that you APPEND as much data as possible, to identify what attributes are MEANINGFUL to identifying a best prospect - one that will convert at 3-4X higher another prospect.
3. What tool do you use for predictive modeling? You don't always need fancy tools for predictive modeling, especially if only have a handful of products/services and a high ticket purchase. I would do the modeling 1X every 2-3 years... Otherwise, it becomes noise and a distraction. The dataset shouldn't change that often. I like CHAID decision-tree modeling.
4. Should you have real-time predictive modeling? You would want predictive modeling / AI built into your platforms real-time WHEN you have lots of segments, thousands of products, repeat purchases throughout the year (more B2C like).
5. Closed-loop tracking - The tough part here is thinking into the future with a tracking strategy, developing naming conventions that will allow for that so that data is uniform. Always recommend a tight UTM strategy- campaigns, medium, source, content, keyword. This should be in use in offline channels too with a simple redirect of an easy link.
6. Appended data - This is critical. The more data you have on your buyers, the better. From firmographics to behavior to contact data to technographics to intent data - any data that would identify their fit as a prospect for you. There are MANY data sources that integrate with MA and those choices will depend on... Many can be real-time if an inbound lead needs that data before scored and distributed to sales. 
7. Segmentation - Be careful about getting platforms that create SO MANY segments through AI that content and relevance becomes hard for your marketing team to keep up with. Each segment deserves it's own buyer journey, content (repurposed from another segment is fine), messaging, and workflows... That creates a lot of work for marketers, and can be a distraction. Consider the highest opportunity segments and create content for THEM.
8. Web Personalization - This is another way to activate your data... based on the segments and data that matters - make each experience as relevant as possible.. Capabilities are in your MA.
9. Progressive Profiling - This is a way to capture data from prospects that just isn't available from 3rd party sources. By prioritizing what that data would be and then REWARDING your targets with content that is high value, you can build up your data set. For example, questions like "Do you currently have a solution for XX" or "Where are you in the process?" (Of course, you also don't want to scare leads off - so when these questions are asked is strategically placed and again, must be rewarded with great content).
10. Reporting- Of course, key to data activation is your reporting. First, sales and marketing alignment on key KPIs is important. I am a fan of marketers looking at cost per sale and profit per sale all-in and by channel (although we know it takes a mix)- and with one aligned attribution method... because this takes into account the whole funnel.  Of course, that doesn't guide decisions, only your performance to goal - you have to break down the reporting from there. The topline exec. results by business units are in BI, and the campaigns and channel performance can be in MA or BI...  If you're optimizing web experience and conversion rates - look at GA (and the set up better be tied to the tracking strategy and naming conventions you identified - so that traffic is grouped in proper channels with proper goal funnels being monitored)... Marketers just have to know what sources to look at for what. A new platform will NOT resolve this IMO...
11. Testing - Testing is what turns data into activation. At any given time, we are testing 3-4 things in market. Whether it's a BIG test in 4-5 markets (with 4-5 control markets) with a going-in trend like the value of TV on the overall volume and CPA (we all know that TV lowers search CPAs because it drives brand awareness - but how much, and what are the alternatives?). Or a SMALL test like what subject line is driving the highest open rate at this stage in the funnel to X persona?

So I believe data activation is most often a PEOPLE and PROCESS issue....

Not to say a new TECHNOLOGY can't help data turn into activation.... I've seen articles that CDPs (Customer Data Platforms) are the answer to our problems here. I don't think that's the case largely.

But I'm hearing a lot about CDPs, that it's the new wave of platforms... Perhaps CDPs CAN help in the following scenarios. It's a question of cost/benefit and based on the business requirements. But in many cases, MA+CRM+BI (plus people and process) is what you need.

When CDPs should be considered:

- Company has thousands of products and repeat purchases where modeling the next purchase in real-time is high-value

- Company has several disparate, legacy home-grown CRMs or other systems that are going to continue to be in use and touch customers

- Company has a SaaS product and uses this behavioral data to target and message and time their campaigns (Boy, this would have been helpful 8 years ago!!)

By the way - I find it odd that CDPs are called "Customer Data Platforms" - there are PROSPECTS that need activated too... I was hoping in this next phase, we as MarTech experts would fix the issue that also plagues CRMs (Customer Relationship Management)...