Microsoft has been talking a lot lately about their AI products that plug into the Dynamics 365 platform, and one of the most popular with Microsoft seems to be Customer Insights. Dynamics Customer Insights (CI) promises the “Customer 360” that we’ve all been chasing for years with integrations, data warehouses, and analytics tools trying to combine various source systems together, so it’s no wonder there is a lot of buzz about it. However, with a product this new (released in April 2019) and this full of promise, customers are bound to be asking where the marketing and reality overlap. Let’s dig into a few key points about CI.
First, CI is the first step in the process of creating a Customer Data Platform (CDP) for your organization. A CDP is a similar idea to a data warehouse or a master data management (MDM) system – a singular place to find all of the data related to your customers. However, unlike an MDM solution, a CDP is mostly read-only – it takes your data in and aggregates it, but it doesn’t make updates to source systems. It’s designed to become a source of truth without impacting any other sources (an MDM will be an active participant with other systems to keep them all on the same page). See Is ‘Customer Insights’ an MDM? for more information.
What Customer Insights offers is a way to create this CDP using the power of artificial intelligence (AI) instead of manually creating mapping logic the way you might with a data warehouse. If you’ve worked on a data warehouse solution, you know that everything is manually defined by developers – this data goes here, these columns match those columns with these rules, etc. For every piece of data, a rule must be defined for where it belongs in the warehouse. With CI, you have some control over those rules – for example, you tell the tool which columns match across data sources so that it understands the relationships in the data – but you don’t need to build logic to match Robert with Bob, Rob, Bert, Bobby, Robby, etc. The AI figures that part out (and you do have control over how exacting you want that matching to be).
Taking some time to understand your data and how it fits together, you can then ingest it into CI through the native connectors (there are over 30, covering most databases, file formats, and some cloud services) and map everything together. Once you’ve defined how you want data to be matched, you run through the Unify process and – presto! – you’ve got customer profiles! Is it really that easy? Actually, yes. The hard part is in understanding your own data well enough to map it effectively and tinker with the matching rules to get results you expect. CI also has capabilities to export data during most of the steps so you can check your work as you go.
After the data is ingested, you have your customer profiles in the CDP. Now you have to turn those insights into action. CI offers a few features to get you on your way. First, there are measures. You can create measures to show you useful indicators about each customer – think things like average annual spend, lifetime spend, number of orders per month, etc. These are calculated from the data you’ve aggregated into the CDP. From the measures (and the rest of your data), you can then build segments of customers that meet certain criteria. For the old-school CRM users among us, this is exactly like building a marketing list – define your criteria and you get back a list of customers that match, in a static or dynamic (updates as data updates) list. The segments can then be exported: natively to Dynamics CE for Sales marketing lists or Dynamics Marketing customer segments, or to Azure Blob storage, or via API calls into CI to retrieve them from any other system or application.
CI also includes additional AI templates you can apply to your CDP data to determine things like churn rate or likelihood to churn. You can use the templates, or you can extend the AI capabilities by building your own models in Azure AI pulling data from the CDP.
Even without the AI insights, having all of your data aggregated into the CDP gives you a way to surface true Customer 360 information inside your business applications. Native connectors in Power BI and Power Apps to the CDP let you build visualizations or interactive apps that display this data to users. Widgets available out of the box for Dynamics 365 CE display a timeline directly on the customer record with all of the activities that CI has consumed. This can give a better picture to a user of what’s actually happening with that customer – not just activities tracked in Dynamics, but perhaps website visits (from web tracking data), orders and payments (from ERP), abandoned shopping carts (from eCommerce), or anything else available.
And from all of this information, you can finally take action: action on marketing campaigns with more granular and applicable customer segments; action on the phone with a customer with a more complete picture of how they are interacting with your organization as a whole; action based on Power BI reports showing aggregate customer activities and behaviors instead of system by system.
Sounds great! Is there a catch? Well, yes and no. Will the product do everything mentioned above? Absolutely. But will it be as smooth as advertised? Well, that depends on the data – what you have available, how clean it is, etc. The AI is good (and is constantly being tweaked to be better) but it isn’t perfect, and if you consider that some data might be hard for a human to figure out what goes where, it can only do so much. You also have to consider what sort of data you actually have available as an organization. CI can’t help you determine your customer churn likelihood if it doesn’t have access to data on orders over time. It can’t tell you when a customer has been on the website if you aren’t tracking that. It can’t link up contacts to their respective accounts if contacts are in one application and accounts are stored in another with nothing to link them. It’s not magic, but honestly, it’s pretty close.
Be sure to subscribe to our blog for more!