Tips to Take Your Predictive Analytics Marketing Strategy to the Next Level

Tips-to-Take-Your-Predictive-Analytics-Marketing-Strategy-to-the-Next-Level.jpg

This post is an excerpt from our Digital Growth Summit event in Sunnyvale.Here are the digital marketing experts who contributed to this blog: Emad Hasan Tips to Take Your Predictive Analytics Marketing Strategy to the Next Level Mihir Korke Tips to Take Your Predictive Analytics Marketing Strategy to the Next Level David Rogers Tips to Take Your Predictive Analytics Marketing Strategy to the Next Level Mayank Johri Tips to Take Your Predictive Analytics Marketing Strategy to the Next Level Jeff Marcoux Tips to Take Your Predictive Analytics Marketing Strategy to the Next Level


 

Tips to Take Your Predictive Analytics Marketing Strategy to the Next Level

Here are some key takeaways from the Predictive Analytics Marketing panel:

  • Avoid the silo mentality
  • Merge, centralize and share data company wide. Pull together teams from different areas to make it the data actionable.

  • Seek out employees who can see the bigger picture
  • When it comes to predictive analytics marketing, the most effective employees are the ones who understand how to get the data from Point A to Point B.

  • Collect data even before you’re ready to use it
  • There are so many tools for collecting data right now. If you wait until you have a clear idea of how to use the data, it won’t be there when you need it.

What challenges do you see with data from a marketing perspective?

There isn’t a common data structure or common way to merge all of the data together. One of the biggest pitfalls that we came across early on was that we had a lot of tools, and we needed to get that insight across the entire company. Instead we had a lot of different groups not talking to each other.

Now we are trying to combine that data and centralize it through our internal data management platform called Cosmos. Our strategy is to suck in all of that data, normalize it and then do actionable intelligence on top of that. It comes down to tools, normalizing the data, and then figuring out how you make it actionable. —Jeff Marcoux

What tools are you using?

From the marketing perspective, we have varied email systems and automation systems we use including our internal product Dynamic Marketing (Microsoft), Marketo, and ExactTarget. We have a great social platform that we are pulling all kinds of data from called Sprinklr, and then we consolidate all of that data into another platform. Then we overlay a ton of machine-learning algorithms over that so we can start to build out the models around that. —Jeff Marcoux

We use a lot of tools that start-ups use. On the descriptive and analytic side, we use Tableau. On the modeling side, we use SAS data modeling. We have a few data scientists on board as well who work on predictive models. —Mihir Korke

Kissmetrics is a great tool for individual data. If you can tie the individual data information back to your data set and work to build your target markets, then that is very powerful. —Dave Rogers

What are some of the key skill sets you need to work in predictive analytics marketing? When hiring new workers what do you look for?

If a data scientist writes the best code but can’t visualize the data, then that is a problem. –Mayank Johri

For me, having the stats background is important, but the other piece that I look for is a social science background in sociology, psychology, something that shows that you are interested in the world. I want to know that you are curious and that you have a natural inclination to figure things out, to peel back the layers. —Dave Rogers

I look for three skills. Firstly, SQL. Then R or Python. It doesn’t matter which, because if you know one you can learn the other quickly. Third, how to visualize the data.

I think that is the biggest challenge. If I have a data scientist on my team who writes the best code but can’t present it and visualize the data, then that is a problem. —Mayank Johri

Share an example of when you used predictive analytics marketing

One approach that we did a bit differently involved machine learning to figure out customer segments. –Emad Hasan

When I was at PayPal, one of the things we were looking at was customer segmentation. We would bring in a marketing agency who would say, “These are their archetypes as customers,” but one approach that we did a bit differently involved machine learning to figure out customer segments. We found a cluster that was buying a lot of women’s products, men’s products, electronics, motorcycles, and we could not figure out what was going on. Eventually we figured out that these guys were essentially couples that were using the same login ID to log in to their PayPal account. —Emad Hasan

What advice would you give to businesses to get ahead of the game in predictive analytics marketing in the next two years?

If you miss something, there is a competitor out who that hasn’t. –Mihir Korke

First of all, are you looking at all of the data that you can gather through the various sources? Are you looking at the people who are visiting your website and collecting all of the information on them beyond the IP address? My advice would be to harness everything that you can and to not leave anything on the table for the moment. If you miss something, there is a competitor out who that hasn’t missed it and is now ahead of you. —Mihir Korke

If you are going to compete with predictive analytics for marketing then take it seriously. –Dave Rogers

Number one, start learning data and connect teams. Do things, get together and start pushing things around. You might have to prove the case at some point, you might not. Number two, if you are going to compete with predictive analytics specifically for marketing then raise your hand and put some money on the table. Take it seriously. Put some effort and energy behind making sure that the data is solid, everyone trusts it, and the connecting points are working. Get some real people doing the work and don’t make it a back burner project. Number three, put teams together. The most successful teams I’ve seen are made up of predictive analytics folks and data science folks talking to BI. BI has the data, the structure, and the warehouse. BI is your friend, so put them in the same room together. I often don’t see that in the beginning, and that is a problem. —Dave Rogers

From a CRM database perspective, how many of you use relational databases in combination with big data types of implementation analytics?

At the end of the day, you need your CRM system to be effective and clean. –Jeff Marcoux

I used to be a firm believer that CRM is the one system. I don’t believe that anymore. I think there is so much data predictive analytics marketing is bringing out that sales couldn’t care about half of it. You need to figure out what are the nuggets that your sales and customer support teams need in CRM that will actually add value and then keep all of the other garbage in another category that you are going to use in insight analysis. It’s not garbage to you, but to them it is just extra fluff and more fields on a CRM field that doesn't help them. At the end of the day, you need your CRM system to be effective and clean, because otherwise it is just a data-input tool as opposed to an actual relationship tool. —Jeff Marcoux