“It’s at the beginning of the journey for CQ. Hopefully it’ll find its feet and then in two years from now, you’ll have better Google results and CQ.”
CQ, a term coined by Ganit CEO, Shivaprasad KT, was something we wanted to learn in our CDO insights series this week.
Shiva is the CEO of Ganit, a seasoned leader in data analytics and business strategy with over two decades of experience. Notably, Shiva served as the Managing Director at Impact Analytics and spent eight years as the Regional Head at Mu Sigma Inc., overseeing key client relationships and contributing to organizational strategy. He also has a successful track record as a Consultant at i2 Technologies, where he led a significant enterprise-scale project for a major tech giant in China and India. His career is marked by a commitment to innovation, talent mentoring, and solving complex business challenges.
AIM: To begin, could you share the story behind the development of the Consumption Quotient (CQ) concept? Were there any specific instances where you saw companies struggle and realized that there was a gap to be filled?
Shivprasad KT – Before I answer the story behind CQ, let me quickly introduce the company. So we are an AI/ML data Science consulting company and we’ve been around for the last six years, a young company, and co-founded by three people. It’s been an amazing journey so far and very gratifying mostly because we’ve been able to help customers realize the value of analytics and engineering.
So CQ is a Consumption Quotient and I’ll talk about why Ganit is excited about it and why Ganit is pioneering this concept and what is the need for this, and because we are pioneering it, Google is unable to give that result today. Hopefully in the next couple of years, we will have done enough in CQ that you’ll get some Google results.
I’ll give you a little bit of a background of how CQ came about. So this actually kind of almost talks about why Ganit was founded and why I was thinking about starting a company like Ganit. So, for me, it’s been almost two decades in the data science & analytics industry. I’ve been dabbling with data for decision making for a very long time. Before starting Ganit what I realized was that a lot of the projects that are commissioned on analytics had primarily three reasons why they were commissioned. Number one was that for example I read this article in HBR, about this new trend in analytics, and can we do something about it? Can you build a model or a dashboard? If the genesis happens just based on the article, I think it’s a recipe for disaster.
The second reason why people want to kind of start a project was because for example my CEO asked me to do this and, if you can do this for me.
And the third was, obviously the real business challenge an organization might face, seeing a decline in Consumer customer base in declining sales and they want to find out why is it happening? What can we do in the future?
So that’s a genuine reason for customers to come to people like us. While the last reason is genuine, the first two reasons were pretty troublesome because if you start with those two premises, what happens is you end up with beautiful looking glamorous dashboards which do nothing for you, but just fill up your table chairs. And you also have a very finely tuned right accuracy print analytics model. But now that is just sitting as a form of a deck or a bespoke tool in some corner in your IT stack.
This leads to what I call the Decision Debt which is, either you’re making a poor decision or worse you’re not making a decision based on this. Some of the solutions that you developed are taken out of context. People take one slide in a bigger deck or a context and they put it into their own deck and then fit their own narrative. All this led to a very staggering statistic which was that about 90% of the analytical solutions that we’re getting created were not getting consumed. It’s very disheartening for me to see it because clearly, there is money and effort behind it. And the data science engineers are working very hard but the customers are not making good decisions. I thought maybe there’s a better way to look at this landscape. Instead of getting excited about just techniques and technology. I thought maybe we should think about the consumption of that solution first and the ability to make a better decision because of that and then get excited about techniques and technology. So I started a company that actually focused on consumption rather than just getting excited about the creation of analytics.
AIM: Can you help us with an example on how companies have struggled because of keeping the conception quotient low. And then how to change that?
Shivprasad KT – I think it’s okay for many people not to understand the concept of consumption quotient today because it’s a relatively new idea, and that’s what we’re trying to educate and pioneer. For example, when you consider a retailer, they typically face a demand forecasting problem. I’ll describe the conventional approach and how Ganit approaches it differently.
In the traditional, non-Ganit approach, when a customer approaches a service provider like us, they might say, “My demand forecast accuracy is 65%. Can you increase it to 75%?” This is a legitimate problem as low forecast accuracy can have real consequences. The service provider would usually take this request and create a project brief, stating that there’s a demand forecasting problem, and the goal is to increase accuracy by 10%. They might set this as the success criteria for the project. Often, this is where the service provider stops.
The Ganit way of doing it is different. First, when we look at demand forecasting, we don’t just see it as a technical problem. We ask, “What business decision does it impact?” If the accuracy increases by 10%, what will it do for your business?
For instance, it might reveal that a significant amount of inventory is sitting in one particular warehouse, leading to inaccurate forecasts. Alternatively, it might uncover understocking in regions where it shouldn’t be happening, creating a different kind of accuracy problem. We guide the customer from the language of forecasting to understanding the broader business implications.
This is the second part of the consumption quotient, where inventory management becomes a business decision. It’s not just about releasing or adequately stocking inventory; it’s also about improving customer service levels.
In the project brief that Ganit creates, we don’t simply aim for a 10% increase in accuracy. We don’t limit ourselves to that. In fact, we ask, “Why only 10%?” Even a 5% improvement in demand forecasting accuracy might result in a $10 million release of inventory. That’s still a significant outcome.
So, we educate the customer about getting excited about a genuine business outcome that directly impacts their bottom line. Whether it’s freeing up inventory or strategically placing it to optimize sales, it’s about understanding the opportunity cost and the financial impact.
This is the fundamental difference between the non-Ganit approach and Ganit’s way of addressing demand forecasting.
AIM: Where do you start the journey to improve your Consumption Quotient for organizations?
Shivprasad KT – That’s a great question and believe it or not, it is the typical problem most organizations face where they have a very strong internal team that does this. The first step here is to involve the right stakeholders. When you look at demand forecasting, it can’t just be a pet project of the analytics COE of an organization; right into the day, the decisions are being made by people in supply chain, in the field, in the operations team. Now you should understand the various decision inventory or the various decisions the different stakeholders would take and hence making a decision inventory of that particular problem and saying that a supply chain person will make these two decisions differently if we had a better forecast, and an operations person would make these three decisions differently if there was a better forecast. Identifying the right stakeholders is the number one step; without that, it becomes a pet project, it becomes a tech project which you should not.
The second one is making sure that you establish successful criteria. You can’t boil the ocean because there is no end to this demand forecasting journey. I’m using the word journey because there is no end goal. So you can keep improving on a regular basis, but you’ve got to have very staggered measurement of your success criteria. What do you call that success in six months after the implementation? Having that success criteria so you know what you are working towards is extremely important.
The third important step here is realizing that these problems are also better solved in partnering with companies like us because we would be seeing these problems across multiple customers, and we bring in the industry expertise of a retail sector. But also, if it solves a similar problem for CPG or a pharma company, we bring in the cross-industry expertise as well, which the customer themself might not have, because of their own industry nuances. So realizing that they’re not alone is the third of the most important steps in this.
AIM: As you mentioned, this would seem fair if the company’s consumption quotient is low. What if the company does not recognise it? How can companies effectively recognize the untapped potential in data analytics to significantly increase profit margins and reduce costs, even when they may not fully grasp the extent of their consumption quotient?
Shivprasad KT – Let’s mathematically define what CQ is. CQ is typically measured on a scale from 0 to 100. Some companies may have zero CQ, while others are slightly above single digits. We measure it precisely. When discussing project outcomes, people often refer to ROI (Return on Investment), but there are two crucial metrics to consider. First, ROI is calculated as your return divided by your investment. For ROI, you need to determine whether you’re looking at a bottom-line return or a top-line return, and then measure the specific ROI you’re interested in. However, in the realm of analytics and AI, where experimentation and the adoption of new technologies are essential, another metric comes into play: Return On Experimentation (ROE). Not all projects have a quantifiable ROI, but that doesn’t mean they don’t bring value to the organization. ROE measures the number of experiments conducted over a specific time period, which should ideally increase regularly. We recommend measuring ROE quarterly, although some industries may opt for six-month or even monthly assessments, depending on their dynamic nature. The key is to combine a solid ROI mindset with an ROE mindset, which is where the journey of measuring CQ truly begins.
AIM: Finally, can you share some key takeaways or advice for companies looking to prioritize and improve their Consumption Quotient and keep the CQ high?
Shivprasad KT – Number one is, if you can establish the kind of business outcome for every analysis. As I talked about, either as an ROI or an ROE, it’s a great starting point. And I think it’s a good way to start your CQ journey. That’s where you’re actually establishing a baseline of your CQ, and then once you know where you are, you can continuously monitor it. The second piece of advice is that for every project, there should be at least one decision that the project enables you to make differently.
So, there should be a decision inventory in the project brief where they state that because of this project, these five decisions will be taken differently. If people do not own up to this decision inventory, again, it will become more of a data project rather than a decision project. Now, you’re getting the drift. Ganit is all about making sure that customers make informed decisions, better decisions than they made yesterday. In this journey, what they need to think about is, while they have a baseline of the decisions they were making earlier, how they can continuously monitor the better decisions they are making.
And a couple more things for people to adopt as best practices: always listen. Open those windows; don’t have a closed house. Closed windows will only result in stagnation. You have to open your windows a little bit so that you can learn from other industries and stay updated with the latest and greatest.
And for every new exciting thing, experiment. It’s okay to fail, as long as you fail fast and learn from it, and maintain a mindset focused on ROE. For example, with Gen AI, I know most of my customers are now asking us, “What are the two or three cool things that you guys can do for us with Gen AI?” So, in the last almost 10 months or so, we have actively been helping our customers create use cases based on Gen AI, which is very exciting. Not all projects, as I said, will have an ROI because of a Gen AI project, but they definitely have an ROE. So, these are some of the things people need to keep in mind when they embark on this journey. More importantly, when they have a mindset of experimentation and continuously building their learning muscle, I think they’ll be doing an okay job.