Drive Massive Growth with Human Machine Customer InsightsAs marketers we're sorting through reams of data to better serve our audience, see trends in our industry, and report on our ROI. Growing as an organization means being able to not only gather this data, but to interpret the various data points. Our hub, SHACK15, was created to build an ecosystem around key players in the field of data science to enable collaborations among those working to advance the field of human machine customer insights.
Developing customer insights, then using them as the core for strategic planning is the key to high ROI and growth, according to a survey by BCG.
And, the problem is very real. Again, according to BCG:
Consumer-facing companies in developed economies have experienced little or no growth since the global recession of 2008 and 2009. As a result, … executive teams are looking outward, trying to spur growth by turning to new sources of customer information. These include “implicit” data such as biometrics, “structured” batches of big data such as online behavior, and “unstructured” data such as social media and call center conversations.
The study of C-suite executives follows up on a similar study published in 2009, with similarly dour outcomes. For instance, only 20% of firms interviewed operated in the top stages when it came to customer insights. And most, are doing little more than traditional market research despite dipping their toes in the water of predictive analytics.
4 stages in customer insights
Take a look at this exhibit from the BCG report and you’ll find the customer insights isn’t an all or nothing marketing activity. In fact, it isn’t just a marketing activity, but involves much of the strategy within an organization. As a firm progresses through the stages, the external rewards get bigger.
I don’t want to go through these stages, but you can get additional information by using the link at the bottom of the image.
Image courtesy of BCG
Transforming your organization into a customer insights powerhouse
In data analytics, we talk about the 3 V’s of data:
But, that’s not all. We also have to worry about the veracity of data. According to some sources, firms spend the bulk of their data resources on simply cleaning the data so it can provide useful insights down the road.
Customer insights come from a variety of sources and take a variety of forms — structured data like from your website and social media analytics, unstructured data from comments, feedback from your sales staff, transcriptions of customer service interactions, etc.
And there’s a lot of it. Every day we create as much data as we did from the start of recorded history to 2008 [source]. This data comes at you at the speed of light. That means no organization can develop sound customer insights without a plan for analyzing this data in near real-time.
Becoming a leader in customer insights requires resources (staff, training, tools) and a commitment from the highest levels within the organization to be effective.
Spending isn’t enough
You can’t just buy your way into becoming a customer insights powerhouse.
As mentioned above, you need a commitment from the C-suite to make it happen. But, there’s more you need. You need a combination strategy involving humans and machines because humans can’t handle the volume and velocity of data entering their system. They also need to take a break, while data streams in 24/7, 365. It doesn’t take a lunch break, a vacation, or celebrate a holiday. In fact, in some organizations, data volume actually increases when the humans in your organization are away.
Meltwater has created a space where those advancing the field of data science can meet and collaborate.
So, what’s a good CMO/ CEO to do?
- Develop programs and processes for machines to acquire and clean data (using AI and machine learning, several vendors offer programs to help clean data)
- Craft algorithms to notify humans when something needs attention. For instance, I created an algorithm using readership data to generate leads for the sales force. Other models might notify managers when a certain pattern of complaints emerge or when a problem needs escalation. Importantly, these models must process unstructured data, as well as structured data.
- These notifications must reach appropriate decision makers electronically whenever and where ever they are. A machine process should monitor responses to ensure nothing slips through the cracks. Hence, when a plane is grounded by mechanical problems, the system notifies engineers and mechanics closest to the problem (and updates the mechanic’s schedules appropriately), automatically replenishes any parts necessary to make the repair or delivers them from another location, as well as notifying executives and gate agents to handle passenger scheduling.
- Create dashboards with critical metrics that require a human evaluation. Use data visualizations for better customer insights and to highlight deviations from standards. For instance, P&G uses a visualization to highlight possible delivery delays to minimize customer impacts by either diverting shipments or informing customers of delays early.
- Ensure marketing staff have the training and technical support to ask the system the right questions on an ad hoc basis. This requires a new skill set that combines traditional marketing concepts, especially understanding of consumer behavior, with business intelligence capabilities to enable predictive modeling to forecast the future. But, more that, it requires marketers who are intuitive when it comes to data and are thus able to ferret out sources underpinning results.
- A strong commitment from upper management that privileges data insights in strategy formation. That means data reports must go directly to the CMO and other C-suite executives rather than lower level managers.
- Build an ecosystem, like Shack15, for those interested in advancing the field of data science to capture human machine customer insights.