Historically, many organizations viewed data and analytics as a support function; one that provides reports, develops complex models, and helps make other functions smarter. However, according to a research report –
- High performing enterprises are 4.6x more likely than underperformers to agree that data is driving their business decisions
- High performance enterprises consistently analyze more than 17 different kinds of data across their analytics apps.
- High performing enterprises are 5.4x more likely than underperformers to primarily use analytics tools to gain strategic insights from Big Data
If business analytics is so strategic, getting business analytics strategy right is of critical importance to many companies. Here is a case study of the approach we followed at Adobe.
Product Management Approach – What is it
Rather than thinking of data and analytics as an after thought, if we think of them as products that internal customers consume and experience, we would think of analytics in a very different way. As consumers, all of us consume data on a daily basis. For example, we all have our favorite news, weather and stock tracking page (Google News, My Yahoo! etc). On most days, many of us as consumers go to these pages to make sense of what is going on in our world. We customize these pages to what is relevant to our world and a lot of times, we make some immediate decisions based on the information (e.g. buy or sell a stock, take along an umbrella if weather forecast so warrants).
If we just substitute news, weather and stock information in this story to information relevant to the company (bookings, revenue, sales pipeline, large deals etc.), decision makers can consume this information first thing in the morning effortlessly and potentially trigger some actions based on it. The experience, when made as seamless as a consumer product, will attract active use of the information, resulting in huge transformation in company’s decision making processes.
A Case Study of Putting this Approach to Work at Adobe
We embarked on executing to the new core data strategy in early 2012 just as Adobe began its business model transition from perpetual software licensing to subscription based business model. This transformation in business model demanded a completely different approach to analytics compared to what it used to be. Like many companies, Adobe had a legacy data warehouse, that pretty much served as a data dump for some core data assets such as financials, order management, procurement, HR etc. Many other data assets were scattered all over the company in disparate data sources. Analyst communities had their own local data warehouses that tried to replicate the same data to build a niche dataset for a specific analytic purpose. Most of the data exchange was batch oriented, with fastest data movement in daily frequency. But since most datasets were assembled after the data jumped multiple hops, latency to latest analytics counted in multiple days and sometimes weeks.
Since similar flavors of analytics were produced by different groups with no governance either around data quality or metric definitions, in many typical scenarios, business groups argued on whose numbers were “right”. A lot of time was spend on reconciling metrics, further delaying decisions and action.
It won’t be surprising if this sounds familiar story to a lot of companies. As bad as this scenario was to support Adobe’s legacy business model, shift to cloud and subscription business would have demanded much higher level of urgency in addressing it. In perpetual business model, customers typically bought software license once a year or two. In subscription model, they have option to cancel and walk away any day and this creates much higher notion of velocity to data and analytics.
The new business analytics strategy, put in motion just as Adobe’s business model transition kicked off, transformed the way company engaged with data.
Product Management Approach Rule #1: Start with Big Vision Statement for Product
Any product development starts with big vision and dream. The vision statement also serves as rallying cry; a statement of intent that communicates break from the past and preparing for changes that would be transformational.
Product Management Approach Rule #2: Make the Vision Specific and Actionable So That Everyone Understands the Direction Being Taken
Great product leaders communicate their vision in a way that is specific and actionable for their teams. It is extremely critical to add some guardrails and design parameters around the vision that may prove abstract for some. These guardrails then become foundational cornerstones in analytics product and customer experience design
Product Management Approach Rule #3: Understand Customer Experience, Customer Segments and Value Proposition
Segmentation, targeting and positioning (STP) is key element of any product strategy. It is important to understand internal customer segments, value of data to them and how they would experience business analytics.
For example, the way a finance analyst would engage with data will be very different from the way CEO or CFO of the company would engage with data. For the former, exploitability and self service of data is crucial as most of her work would be focused around explaining causation of business events. Senior executives on the other hand, would directly engage with data by tracking key business metrics with limited self service expectations. They would perhaps expect mobility – accessing data on the go and at will.
A powerful tool to develop clarity around segmentation, targeting and positioning is the Business Model Canvas. This canvas helps you quickly build what personas of customers you would target and what kind of experience you should plan for them. It will also help clarify what kind of analytics are consumed directly versus indirectly – i.e. produced by business analysts.
Product Management Approach Rule #4: Brand Your Product to Reflect Value Proposition
One of the problems we attempted to solve was – how to raise trust in numbers. Because of the prior fragmentation of data architecture and proliferation of poorly managed standalone analytics databases, trust in metrics and numbers produced by analysts was low. While the business analytics strategy addressed the underlying problems, the challenge was – how do we communicate a brand of trust.
To address this issue, we created our own product brand – Adobe Dash. The idea of this brand was to reflect similar values to – say Intel Inside branding – i.e. if one saw this brand on analytics, the output was backed with solid data architecture, standardized metrics definitions, data quality exceptions monitoring etc.
This changed the attitude of the company towards analytics very quickly. We quickly started our user base demand analytics come from Adobe Dash as they realized trust in numbers saves a lot of unproductive discussions around data quality and keeps focus on informed decision making.
Product Management Approach Rule #5: Base Technology and Platform Decisions on Product Strategy
Platform and technology decisions ought to follow product strategy. For example, “business intelligence at the speed of thought” paradigm which led to two key tenets – fast user experience and low data latency (or near real time analytics). This led us to choose an in-memory columnar database that stored data at lowest granularity in most cases while allowing aggregation in real time with millisecond response time. This also allowed us to provide the same consistent numbers across a host of BI – user experience tools as we consciously made choice not to cache data in these tools. We also selected user experience tools that deliver stellar mobile experience for our executive segment.
Speaking of user experience, it pays to invest in user experience designer to make your analytics products look beautiful. User experience has been used as a massive differentiating factor by likes of Apple. If you have coolness factor around analytics, you will have more likelihood of wider acceptance across company, especially at senior management level.
Product Management Approach Rule #6: Drive, Monitor and Measure Adoption
If you had real product sold in marketplace, customers would vote from their pockets to communicate whether you are successful or otherwise. For internal analytics, your users would vote by using your analytics products. Therefore, it is very important to track usage trends and focus on any areas of weakness.
We specifically tracked usage by Analytical Product (e.g. Sales Performance Dashboard), by organizational groups, by organizational levels, by frequency of use etc. and derived interesting insights that helped us fine tune these products.