Selling knowledge graphs-The exciting journey towards the unknownSudhanwa Rajurkar
“The Journey is the Destination“… I could not find the Zen proverb fitting to a technology solution so well till now as it fits Knowledge Graphs.
I won’t get into what is a Knowledge graph, because there are tons of articles on the web that explain it in easy ways. The Internet has plenty of information about the advantages of KGs, challenges in implementing KGs.
I would rather like to share our experiences of the real-world challenges that we faced when selling knowledge graphs as solutions to real-world problems, and how did we go about these challenges.
With the overwhelming growth of digital data and information overload, Knowledge Graph-based solutions are not a competitive edge anymore.
It’s a must-have to help businesses to achieve their Moonshots!
What is more interesting about the Knowledge graph is that it is a journey, that needs to be undertaken by the business stakeholders and the Data Scientists, collaboratively and knowingly. Embarking on this journey needs a great deal of data storytelling and ROI mapping.
But, It’s important to know that the benefits of knowledge graphs are realized along the journey; benefits are manifold along the journey, but not all are known upfront.
All of this makes Knowledge Graphs a difficult sell to the top management. So how do we convince the top management to embark on this journey?
We explain the power of inferences on inter-connected data through relevant data stories and examples, which make them want those inferences, which today can be done only cognitively. Imagine you are talking of millions/billions of such inter-connected data points, it’s a humanly impossible task to come up with accurate inferences for solving complex business problems, and quickly. We Establish the importance of scale of connected data, speed of inferences for decision making.
What really helps is asking them to go through an exercise, where we ask them to define their business requirements, expectations from the solution in the following ways
- Short Term expectations – Identify must-have features that will bring them parity with competition
- Long Term expectations – Build USPs for their business
- Vision – We ask them to dream/imagine what they wish to have in their solution, which they think may not be possible in the near future. It’s their moonshot!
This helps us create a data model, connections, entities, metadata with the long term intent, and also helps us deliver on near term expectations
We then collaboratively choose a few expectations (based on business importance) and create a short first release plan to deliver these expectations.
The Secret Sauce – It may be sounding too simple, but it’s easier said than done. With the advent of new technology options for implementing Graph storages, Knowledge layers within the KGs have increased since last few years; but the secret sauce is actually in being able to formulate the KG solution to utilize the business data, model it into an interconnected graph that can yield unseen and unknown inferences to satisfy the given expectations.
In this journey, we are not only able to deliver these near-term expectations; but also uncover new paths in the Knowledge graph to deliver on much more than expected. When the data insights come in, suddenly we can see new paths taking shape.
One such legendary story is famous as ‘Diapers and Beers’ with Walmart. When Walmart deployed KG for their stores, they were astonished to see that the purchase of Beers went up on Friday when Men bought Baby Diapers. (Given the impact of the new father planning to enjoy with friends on weekend). When they moved these two items together, the sale for both increased.
These kinds of discoveries can not be predicted upfront as a ‘destination’ while analyzing the data, but that’s where the Magic of Knowledge Graph can lead to.
This approach helps convert the business aspirations to outcomes, in the early stages of the journey, and keeps the excitement going for the further expectations and unseen value which is waiting to be discovered.
At Coreview, we are using this approach with multiple customers, and it is in different stages of the journey with each of them. In these implementations, we are seeing mixed success and interesting inferences. It is already showing a positive impact on their business growth. With further tweaking, I am sure we will be able to delight our customers. I am sure we will help them reach at least the stars, if not the moon.
So, What’s your Moonshot?