Self Service Analytics. The industry has been promising this (and failing) for decades. I know it has, because I am essentially a data dinosaur. I started my career implementing Executive Information Systems (EIS) and Decision Support Systems (DSS) from Pilot Software back in the ‘90’s (yes - 90’s as in 1990’s!). These systems promised to give business leaders all of the data and information they would need to ask questions and make better decisions. Unfortunately, these tools were quite rigid. In reality, they were hard coded dashboards that offered limited functionality.
As technology evolved, companies started selling OLAP (Online Analytical Processing) tools to allow easier interaction with data by modeling it in multidimensional terms. I was working at Brio Technology during the height of the OLAP wars (yes, this was a real thing and no it was not as exciting as the phrase may imply). Whether the underlying software was based on DOLAP (Desktop OLAP), ROLAP (Relational OLAP) or HOLAP (Hybrid OLAP) technology, they still did not fulfill the self service promise. Business users could do things like drill into details but they could only ask questions where the answers could be easily derived from the underlying data model. They were left with dashboards that only answered questions that had been asked previously.
Later, I went to work for Tableau Software which focused on giving business users the ability to use and understand data. Tableau took a different approach and focused on ease of use and visualizing data in ways that made it easy for the business to understand. However, like its predecessors, eventually data teams used the tool to build dashboards (beautiful ones!) and business users only interacted with those - instead of being able to answer ad hoc questions on their own as they arose.
That was a long winded way to get to my actual point: self service analytics has not come to fruition. It has not been possible for two major reasons:
BI and Analytic tools rely on a table, schema or defined model. Data engineers usually create these by cleansing and joining different data sources. With the number of data sources increasing and the influx of Saas applications, the process of creating a clean data model that joins data from all of these disparate sources is extremely laborious and difficult to maintain. Whenever something changes or a new vendor is onboarded, data engineering teams have to update, test and validate all of the datasets they have created. Additionally, these datasets are constructed from defined requirements. As in the past, the datasets are created to answer specific sets of questions. When users have new questions, they have to go back to data engineering for them to modify or create new datasets. By the time the data is ready to query, the business may have moved on to new questions. Business is still reliant on a data team to create the dataset they need to answer a question, so they are never able to have true self service analytics.
There are lots of great tools with easy visual interfaces. However these tools always look at things in terms of tables and columns and keys or alternatively they use terms like dimensions and facts or metrics. If a business person has a question about repeat purchases and they have access to an orders table, they quickly become confused about which columns to select and how to manipulate the data to get the answers they want. Data is presented in its storage format, not in terms of business processes and logic.
To finally allow self service analytics, we must think differently about how we model data, especially when companies now have so many different data sources (according to Blissfully, the average company uses 137 unique SaaS applications on average). Data engineering can not keep creating different combinations of joined data sources with different granularity.
Narrator has created a unique platform which successfully addresses the challenges that have plagued historical attempts at enabling self service analytics. The secret to the solution is in how data is modeled. Businesses need to start modeling their data in terms of the business activities that actually create the data. The activities need to be defined from just the source that stores that activity data without having to create complicated joins with other systems. For example, if I send an email through my marketing platform, that information is written to the marketing system. I can easily define a “send email” activity with the pertinent data from just the marketing system. All of a business’s activities can be simply defined from just their individual source of truth.
Once defined, if using a common structure, all of these business activities can be combined into a single table providing a true “Customer 360” view. (Obviously, the system deals with identity resolution, duplication and a host of other extremely difficult things!) This single table not only enables instant insight to every activity a customer has touched across all systems, but also allows anyone to answer any question without having to involve data engineering teams. Business users can ask questions without choosing tables and columns, instead, by choosing activities. To see if customers who attended a webinar converted to purchase a product within 30 days, they would choose the two activities (attended webinar and ordered product) which were originally captured in different systems. This would provide the data for each customer who attended the webinar and the details of their order - if they purchased within 30 days of the webinar attendance. The raw data can be grouped, filtered and charted. To really empower the business, this data is accessible for other tools and integration. Business users can use the data in any manner they need, knowing that it is coming from the source of truth.
Narrator’s data platform allows a company to quickly and easily create a customer 360 table that not only gives easy access to any customer journey but also allows anyone to ask any question in business terms. After all these years, Narrator has finally found a way to quickly enable self service analytics across the entire enterprise.