A Dashboard Everyone Ignored
Three years ago, I was called into a regional online retail company to diagnose a growth plateau. Despite a 40 percent increase in their marketing spend over the prior two quarters, they had flat revenue growth for those same periods. Throughout the organization, there was frustration at the leadership level while the marketing department was defensive, and the sales department was placing blame elsewhere in the organization.
The first thing I requested was to see their analytics set-up.
They had GA (Google Analytics) set up; they had a Shopify dashboard; they had a custom report created by the operations manager that required 4 hours of work to complete each Friday; they had a dashboard for paid social activity from their agency; and finally, they had a spreadsheet with 17 tabs, using multiple colors to indicate how someone who had extensive knowledge of it (and no other employee did), would be able to interpret it.
Data? They had plenty. They had terabytes worth of behavioral information, transactional information, customer history, advertising performance data, etc.; however, the data they had was located within distinct systems that did not speak the same language or tell the same story.
Decisions weren’t being driven by anything but gut feel or anecdotal sales feedback and then the data was being used retrospectively to validate already made decisions.
Is this sounding familiar to you?
That’s true for most mid-market companies today. They are rich in data but poor in insight. It’s the gap between these two states that is costing them in the form of missed pivots, poor decisions, and handing over opportunities to competitors who have figured it out first.
Data vs Intelligence
Let me straighten something out here from the very beginning, Data should be viewed as raw materials and Insight should be viewed as finished products. However, the majority of organizations have mastered the acquisition of raw materials and have severely underinvested in their refinery.
Think about crude oil – you cannot operate a car with crude oil; instead, crude oil must undergo a series of processes (refining) before turning into a usable product (gasoline) to operate a vehicle. Similarly, data must also undergo refining (processing) before converting it into something usable (intelligence).
For example, data is the time of a customer’s session on your pricing page, but the insight might be that customers who spend more than 90 seconds on this page convert at three times the rate of those who do not spend this amount of time on your pricing page; therefore, this insight would provide you the means to create a sales trigger based upon this behavior. One is just a number; however, the other one allows you to develop and act upon a decision.
According to research conducted by McKinsey, organizations that are in the top quartile when it comes to data-based decision making are twenty-three times more likely to acquire customers, six times more likely to keep customers and nineteen times more likely to be profitable than organizations that are in the bottom quartile. These are not insignificant advantages; they are clear and critical advantages based on poverty versus wealth.
The real question is not if data matters, but does your organization have the technical and cultural capabilities to turn data into a decision as fast as business operates?
How Real-Time Analysis is Changing Everything
Most discussions around data analytics get stuck in the past due to the nature of how traditional business intelligence is built. Business Intelligence was historically built to follow a reporting cycle — weekly dashboards, monthly reviews, quarterly deep-dives; you looked in the rear-view mirror to see what happened, you nodded your head and created an action plan for the next quarter.
This is no longer a viable model because markets do not operate on quarterly cycles any longer.
An individual’s buying behaviour may change within hours. The price of a competitor may change overnight. A supply chain interruption may start real time. A social media event can create a spike in demand, or create a decline in demand, before your weekly report will even be produced.
The concept of real-time analytics goes beyond simply providing speedier reports to now represent an entirely different way of operating. You are now able to base your choices on ‘ what is happening now’, rather than relying on data that was compiled 30 days ago and might not reflect what is currently going on in your business environment.
Many examples can be given to provide clarity around this concept; one of the best examples is Netflix. A company that regularly gets media coverage due to their effective use of technology and smart business practices, Netflix uses a recommendation engine that constantly analyzes viewing behaviours in real time, meaning as viewers watch programming, the system can update its recommendations to reflect the viewing session. However, what is seldom discussed regarding the Netflix recommendation engine, is how Netflix leverages real-time viewing data for operational purposes, specifically in relation to content production decisions. Through identifying the types of genres, themes and story structures that lead to successful viewer completion rates, Netflix then uses that information to develop its content supply chain.
This is not unlike an oil refinery that takes raw oil (viewing data) and produces unleaded gasoline (revenue-generating decisions) by processing the raw materials through a series of steps to produce the final product.
Layer 1 — Data Integration (Collection Problem)
The first layer involves ensuring that the data source(s), such as CRM, marketing automation, and website analytics, can communicate with each other to produce a single version of truth rather than multiple “parallel universes.” This means there is either a requirement to build or acquire a correct data pipeline (Fivetran + Snowflake + dbt + Looker for an enterprise-level solution or something more basic for small companies) in order for data from different locations to be integrated and produce a single set of results. In order for there to be complete business insight through real-time analytics, access and integration of data from multiple disparate locations must occur, which is the basis of the final layer of the infrastructure.
The value of raw integrated data “insight” is not as valuable; therefore, analytics architecture provides insight. For example, analytics architecture provides metrics driven by business objectives and metrics that lead to an indication of performance change prior to any change occurring (e.g., revenue).
Vanity metrics are metrics that can be tracked quickly; therefore, most dashboards are built using vanity metrics (e.g., sessions, impressions, follower counts, etc.). The metrics that drive business decision-making, however are usually not reported on or tracked (e.g., customer acquisition cost by Segment, cohort retention curves, Product engagement depth, Revenue per source).
To determine the best progression to define your metric hierarchy prior to designing any dashboards—what are the decisions you need to make? What updates to your decision would change your decision to generate the data—working backwards from the decision to the originating data source.
Analytics investments stall when it comes to activation. You have integrated data, built intelligent dashboards, and now what?
Gaining insights from data is labor-intensive and costly; the only way to realize the value of the refinery is if the output actually produces fuel. Therefore, we must build real-time analytics into operational workflows so that they are part of the day-to-day operational activities, rather than just reviewed in quarterly business reviews.
From a practical perspective, “real-time analytics” means using automated alerts when key performance indicators (KPIs) exceed an acceptable limit. It also means that sales teams will have access to real-time lead scoring and be able to update their priorities based on buyer behavior. It will allow marketing budgets to automatically shift to channels that show positive trending performance. And provide updated pricing models to help achieve profit optimization based on inventory and demand signals.
Activation is the point in time where data can start generating revenue. All of the prior steps are just supporting infrastructure.
Business Transformation and the Unlocked Potential of an Operational Model to Drive Profitable Revenue
To illustrate what I mean and support my point on activating insights from an operational perspective, let’s go back to the case of the e-commerce retailer I previously referenced on page 1.
We ultimately were able to help them realize significant revenue enhancement opportunities by applying enhanced analytical processes that integrated the company’s advertising data with the customer relationship management (CRM) data and the customer’s purchasing history; thus providing the organization with a comprehensive view of a customer’s entire purchasing lifetime value (PLTV) based upon where the customer was originally acquired through a previous channel.
Through this data mining exercise, we found that customers who were acquired through one specific mid-funnel (content) channel had a PLTV that was 2.4 times more than the average PLTV for all customers acquired through one of the company’s top-performing paid social advertisement campaigns, even though the dollar volume generated from the mid-funnel content channel was approximately only 1/3 of the total volume of all the paid advertisement campaigns across all social media platforms.
They were scaling the wrong Acquisition Source for 18 months. Not due to a shortage of data. Because no one had connected the dots.
Once they were able to see that signal clearly, and monitor it in real-time, they shifted 35% of their paid budget to nurturing the content channel. In two quarters, average Customer LTV across their entire acquisition portfolio increased by 28%. The amount spent on marketing remained the same. The output was considerably different.
They were able to make that pivot because they had finally built the refinery. The raw numbers had been there the whole time.
Where Most Organizations Go Wrong
I’ll be direct about the most common failure mode I see: organizations dump a lot of money into Data tools and then call it a Data Strategy. Organizations purchase a BI platform, designate someone to oversee with belief that data will create insight.
It does not happen this way. Tools do not create insight; Analytical thinking will create insight. When Critical Thinking exists, Tools are simply faster/scalable for Analytical Thinking.
Another common failure goes to treating Data Analytics as an IT Function versus Strategic.
When decisions regarding data infrastructure are made exclusively by a technical team with no input from business leadership, you’ll often end up with a technical system that provides answers to problems that need addressing but are not currently being addressed.
The most successful analytical operations I’ve seen all have something in common; they provide a means to make connection between a company’s data capability and executive-level decision-making. As a result, there is no layering through 3 levels of middle management before an insight reaches a decision-maker, instead it goes straight to where the decision is made.
Teams like Yeasinarena have developed this as their operating model to help organizations build their data analytics operations. To do this they don’t only stop at implementing dashboards or building connections between data sources but also help design the complete path from raw data to business decision. They operate at the crossroad between technical structures and strategic intelligence, and in doing so create real value in the organization. Most organizations do not recognize the importance of creating both simultaneously until they have attempted to do so on their own.
The Refinery Mindset
One of the biggest shifts in thinking is to stop viewing data as something that needs to be collected and reported, but rather to see it as something that should be refined and utilized for action. Every piece of data collected by an organization has the potential to provide intelligence about:
- Customer behavior and patterns
- Market response
- Operational efficiencies
- Competitive landscape
All of this data is considered “raw materials,” which can be transformed into decisions regarding revenue generation.
The companies that will lead their respective industries in the next five years will not be those with more data than anyone else, but those with the best refineries to refine that raw material into valuable insights.
- Examples of the critical components of an organization’s data infrastructure include: real-time analytics, an integrated data infrastructure, and insight activation integrated into daily operations. Collectively these tools can turn numbers into growth.
If you have more data than you are effectively using, or if you would benefit from making better decisions based on the analytics you currently have, you should conduct an audit to identify the gap between what you are capturing and what is being utilized for decision-making. If you’re looking for a strategic partner to help close this gap, Yeasinarena specializes in assisting organizations have these conversations.