Microsoft's Jeff Marcoux: Predictive Marketing: What Does it Mean for the Modern Marketer? Live

Jeff Marcoux, CMO Lead, Worldwide Enterprise Marketing at Microsoft Corp. and board member of the Internet Marketing Association
By Jeff Marcoux, CMO Lead, Worldwide Enterprise Marketing at Microsoft Corp.

The world for marketers is a constantly changing landscape of new technologies and platforms to target and market to our respective audiences.

Recently, Sirius Decisions published a blog on marketing automation which showed that adoption of marketing automation software is still just ramping up, but quickly approaching the inflection point in the adoption curve, the hockey stick. Add onto that all of the new data sources that marketers have, including marketing automation analytics, and you end up with a mess of metrics that doesn’t offer much value in its unprocessed state and at best provides a backward looking view.

What good are likes, retweets, shares, opens, and clicks if they don’t stitch together to create an amazing customer experience and personalized story?

I recently finishing teaching a class for UC Irvine on transmedia marketing and it got me thinking about how the face of marketing is in the process of changing forever – the driving force -- machine learning and predictive marketing.

What is Predictive Marketing? 

Many are throwing it around as a new buzz word, but their platforms are not “true” predictive platforms. What do I mean by that? Well, let’s start with a common confusion between forecasting and predictive. While they both are focused on guessing at something in the future, forecasting takes historical data and makes a best guess at what the future will hold -- weather, how many Apple Watches need to be made, etc.

"Predictive marketing is a way to leverage data in a way that adds value and positive impact to your end users like never before."

Predictive analytics, on the other hand, takes into account behavioral and contextual data as well -- customer behavior, the psychological side, social data, etc.
Add into this machine learning and you create a predictive model that learns and adapts over time to optimize itself. Why is this important? Because each of them have important uses for businesses, but predictive is going to be a game changer for marketers. Predictive marketing is a way to leverage data in a way that adds value and positive impact to your end users like never before.

Why Now?

In the past, predictive and forecasting have required extensive statistical backgrounds and have had fairly limited data sets to draw from. Additionally, the computing power and cost for access to machine learning software was astronomical (often in the price range of $1 million a seat) and unaffordable for businesses.

Today though, these technologies are much simpler and affordable for businesses and end users to take advantage of. You can do very simple forecasting with a simple click and drag for any data set, data visualization is no longer limited to just a few in an organization, and machine learning is becoming accessible to small and mid-size businesses through products like Microsoft’s Azure Machine learning. This is a recipe for a data driven, innovative, growth hacker culture -- the future of marketing.

Already, predictive analytics are already being leveraged in some amazing ways:

  • Amazon’s recommendation engine: Although this could easily be taken to the next level away from "people who bought this often bought [insert product]", to incorporating social graph data to further personalize recommendations. E.g. I see you have been sharing about recently having your baby and complaining that you can’t find any newborn diapers. The engine could serve up a subscription offer for newborn diapers. It's a very simple example, but you get the idea -- incorporate social context to a recommendation engine could make it personal and take it to the next level.

  • Search engine marketing bid optimization: SEM analysts beware -- your days are numbered. Eventually, this will likely be directly linked to buying media as well.

  • London Underground, Internet of Things: The London Underground has implemented sensors all over their system to enable better security and proactive maintenance. This is all about enhancing the customer experience so trains are on-time, escalators work and ticket counters are never down, making it more efficient and easier for customers to use the Underground.

  • Inventory forecasting: This goes beyond simply what do we expect demand to be at XYZ store to incorporating weather forecast, seasonality, sports teams, political climate, culture, etc. in an area to ensure that when consumers go to a store, they will always have what you want.

  • Fraud detection: Several businesses are leveraging predictive analytics to identify patterns that are fraud signals, forecasting inventory to ensure stores have what consumers want when they go in to purchase.

Possibilities With Predictive Marketing

This is just scratching the surface of how predictive is being used and what the possibilities are.
If we truly look at all of the potential data sources that predictive marketing can tap into, the applications are endless. Data sources like social sentiment, weather, seasonality, political climate, the Internet of things, wearables, the social graph, website analytics, CRM data, email/marketing automation data and the list goes on and on. There are five areas where I think that predictive marketing is going to have major impact in the next five to 10 years:
  1. Automate optimization processes: We are already seeing this start to occur for early adopters of the technology. Things like website and email A/B testing, SEM, pricing and inventory will all become automated processes for businesses as the analysis can be done via machine learning and constantly adapting faster than a human could.

  2. Lead generation & content: Startups like Everstring are already beginning to do this, but imagine where you have a clear understanding of your best customers and types of companies you can sell into. Your predictive engine is constantly trolling the web for signals that identify new target buyers before they know they are even interested in your product. Additionally, predictive marketing could go as far as to serve up the best possible content to your sales team based on the deals they are currently working before they even know they need it -- industry, product interest, size of business, etc. This all is automated to make the handoff from marketing to sales seamless.

  3. Taking customer loyalty to a whole new level: Marketers could leverage predictive analytics to know what their customers are using, how they are using it, and identifying signals that they may be looking to leave your product/platform. With data like this you could predict the best types of information your customers need to succeed and fall in love with your products. Imagine making call centers a thing of the past because you are able to proactively reach out and resolve issues before the ever escalate.

  4. Predicting campaign success before they ever launch: Imagine if you could run models against campaign ideas before you invest and execute? Know what will work, what wont, and the ROI? Which channels and audiences you should target, how much money to invest and where, what content you should create, and beyond. This is the future of marketing in a nutshell. Where we only execute things that have been tested and we are able to maximize the return on every dollar spent in marketing.

  5. Predictive personalization: This is one area where I see the most potential benefit for marketers. We are all obsessed with the customer experience and the journey that people go through as they engage with our brands. However, every person is different and interacts with us differently, why shouldn’t our businesses treat them all as the individuals they are? This is where predictive personalization comes in: Every single person should have a different engagement with your business in context of what they care about, all with the goal of moving them forward in your sales cycle. Imagine you have three different people from a business looking at your product: CIO, CMO & Director of Demand Generation. Each of them are looking for different things and with the data sets listed above, businesses could easily curate the right content for each of those audiences on the website before they ever have any interaction with you. All aligned to their interest, needs and stage in the sales cycle without you having to do anything. No longer do they have to hunt for what they are looking for but it is served up in a compelling and unique way. Predictive marketing is how brands will have to differentiate in the very near future. Marketing automation and simple website personalization will no longer be enough to create a differentiated customer experience, lead generation and identification will become much easier for marketers, process optimization will go the way of machine learning. 

    "Data is the future but we have to find significance in that data and turn that insight into predictive engines that will enable us to exceed our customers’ expectations and drive marketing into the future."

    Data is the future, but we have to find significance in that data and turn that insight into predictive engines that will enable us to exceed our customers’ expectations and drive marketing into the future.

    I will be discussing this topic more in detail at the Internet Marketing Association Conference, Sept. 24-26 in Las Vegas: For more on Autonomous Marketing, see my blog on The Human Side of Autonomous Marketing. Feel free to contact me or add me to your LinkedIn:

    Follow Engage on Twitter and bookmark the Engage Magazine website. Click here to see how ScribbleLive's content engagement platform can help your company achieve its business goals

    Related Links


    Font Size
    Viewer Comments
    Translate posts and comments.
    Powered by ScribbleLive Content Marketing Software Platform