Rajan Sethuraman is a highly accomplished professional and an alumnus of IIM Calcutta, where he completed his Master of Business Administration (MBA) in Business Administration, Management, and Operations during 1993-1995. With a strong background in management consulting, Rajan brings extensive expertise to his role as the Chief Executive Officer at LatentView Analytics, a trusted analytics partner to some of the world’s most renowned brands.
Throughout his career, Rajan has showcased his proficiency in various domains, including Global Delivery, IT Strategy, Management, Business Development, and Business Process Improvement. His leadership skills and strategic acumen have contributed to his success as a CEO, making him a valuable asset in the field. Rajan’s track record reflects his dedication to delivering excellence and driving growth in the organizations he serves.
In this week’s CDO Insights, we interviewed Rajan Sethuraman to gain insights on balancing innovation and meeting client needs effectively, as his extensive experience and strategic leadership at LatentView Analytics align with the topic’s relevance in the analytics industry.
AIM: How do you view the relationship between innovation and client needs, and what experiences have shaped your perspective on striking the right balance between these two priorities in your professional career?
Rajan Sethuraman: One of the things we notice quite a bit, is that while there is always an interest in trying out new things and being innovative, there is also this need for being grounded in the business reality of our client’s current context. Especially in a situation where budgets are tight and you’re operating under a source constraint, there is always the need to make sure that every dollar you spend impacts the P&L. In that context, it becomes crucial that while you are being innovative, you’re also doing something that will have a real impact on the bottom line or top line of the company. The decibel levels around generative AI and large language models are at a peak right now, especially because of the democratization that has happened in recent times. Through ChatGPT, everybody is now familiar with it, and there’s also a fear of missing out creeping in – “If I’m not doing something in LLM, am I missing out?”
However, I think in the last two or three months, I’m starting to see some changes, at least in terms of client perception. Instead of just saying, “Let’s get something underway and let’s get something started,” they’re asking questions about how they can make meaningful use of generative AI or large language models or any other emerging technology. In this resource constraint scenario, we were hoping that the second half of this calendar year will be more relaxed from a budgetary standpoint. But that is not panning out. Clients and organizations are still walking a tightrope when it comes to requirements.
One of the things we do when it comes to innovation and new ideas, is to try and sit down with the client and help them think through what specific metric they are trying to impact. It could be a financial, process or an operating metric. We peel the onion a bit more to get clarity. We let them know the initiative we are undertaking, while attempting to use generative AI or LLM to improve its performance but focus more on the key metrics we are targeting. We ask them if they are aligned with this, and if these metrics are important in their current context. If yes, we initiate a pilot and the POC. However, most of such initiatives are still in the pilot POC stages; I don’t think there are too many full-fledged ones. You will find something in customer experience, customer interaction, and so on. That is where NLP and NLG technologies have also been used in the past, and GenAI is helpful. But when it comes to hard data analysis, generating insights from a ton of available data, people are still figuring out how to marry the hard science, the math, and the modeling with the GenAI capability.
AIM: Why, particularly in technology, do clients often seek to implement solutions like Generative AI without a clear problem statement or use case, despite having specific business metrics and objectives in mind?
Rajan Sethuraman: My hypothesis is that it is because of the pace at which technology is disrupted. In the last couple of decades, we have observed the life cycle – the introduction of a new technology, the hype cycle, getting to real impact and maturity, and then getting replaced by something else – is getting shortened with every passing quarter. New technologies come up, and you need to be quick in how you understand it, how it can impact your business, and quickly start leveraging it. And obviously, the underlying fear is that if I’m not on it, I’m missing out on competitive advantage. And this could be a real competitive advantage. Whether it is in terms of innovation, products, and services you can come up with and put in front of your clients, or whether it is managing costs, a larger stake, your supply chain, and your employees – all these areas are being impacted by technology. The new disruptions are happening, and the disruption cycle itself is shortening. So, if you take too much time to understand and think through and use more traditional methods, then the danger is that you might have missed one entire wave. Sometimes, people think they can leapfrog and get onto the next wave if they miss one wave. But, if you didn’t understand and capitalize on the previous wave, then how are you confident you’ll do it in the next wave? So, it’s like running on the treadmill. You must keep running and move fast even to stay at the same pace. So, my hypothesis is that the rate of change and disruption is resulting in a mind shift that is happening. Clients want to say, “What can I do now? I know that this is new technology. Maybe it is not even mature, and people have not understood it clearly, but can I be the first to do that? Can I be disruptive?”
AIM: How do you view the long-term implications of persistent technology adoption without well-defined use cases in industries moving towards digitization? Do you advise clients to embrace this trend or advocate for a more strategic evaluation of technology adoption?
Rajan Sethuraman: It’s dependent on context to some extent. If you have a solid business model with a good amount of protection, entry barriers, and modes, and you are already differentiated in what you do, the products and services you offer, then you are not looking for the next new thing to stay relevant and differentiated. But you also notice that not all business models are along those lines, especially in the last ten years with the burgeoning digital economy, the Internet of Things, and so on. There are several business models that are in existence today that capitalize on, “Am I on the latest, from a technology and an innovation standpoint, and can I quickly ride that hype cycle or the maturity cycle, and then move on to the next thing that has increased significantly?” If you’re a company that is in a space where you need technology disruption to stay relevant, then you just have to go with it.
It’s not a question of healthy and unhealthy as you can look at it from multiple perspectives – from a shareholder, employee, or technology standpoint. But there is also a question of relevance – is this necessary? Is it relevant to what I’m doing? I think that might be driving the action more than whether it is healthy. Obviously, most organizations talk about a foot in today and a foot in tomorrow. There is the pharmaceutical industry, for example, where the typical life cycle of a new product introduction and drug discovery processes can take a long time. But even in those spaces, you now see the technology disruption. Image Analytics is a big play right in the new brand discovery. The latest model around protein folding, for example, is AlphaFold. These are significant in the scheme of things today. Everybody is looking at how we can shorten product development cycles. I guess it’s an ever-changing “Ooka” world, as they call it. A little bit of this fast-paced change is inevitable.
AIM: When balancing innovation against immediate needs, do you have a specific framework or approach to guide client expectations? Can you share insights into how you manage this dynamic effectively?
Rajan Sethuraman: I’ll talk about a couple of things that we do today, which we have found to be practical and sensible when having conversations with clients. For ongoing engagements, when working with a stakeholder, we have mechanisms like the Monthly Business Review and the Quarterly Business Review that focus on the initiatives currently in play and what is the result and impact of those initiatives. Every month, we look back at what we were trying to do with the analytics initiative and what metric we were trying to impact, and what decision processes did we make a difference to. There is a lookback that we do in every monthly and quarterly business review and identify how much we could move the needle on the metrics in question. Because only if you understand that, you can look at what’s next. What we also do in the same MBRs and QBRs is to table a few new ideas and innovations. For instance, “Here are three things you should consider for the next quarter or year”. It could be about extending an initiative into new geographies or products or markets, or it could be a new algorithmic approach to solving a problem more effectively. It could be a completely new technology disruption like GenAI. In every meeting, we spend an equal measure of time – the first half looking back and the second looking forward to what’s next.
This mechanism has been serving us effectively. We survey with our clients every quarter to understand how we are doing. We call this the Voice of Customer survey, and get formal feedback from every stakeholder, and one of the critical questions that we ask them is, “Do you see us as an execution partner, a thought partner, knowledge partner or a consulting partner?” A consulting partner is where even if the client is not clear about the requirement, or if they’re asking high-level questions like “What is the next thing I should be worried about, or that I should be capitalizing on?”, we help them through the process. It’s heartening to note that in the last six quarters, we’ve been trending up on those questions, where our clients appreciate that we are bringing those ideas to the table. They may not always take the ideas and run with them immediately, but it’s food for thought. It’s new ideas that are being tabled.
The other construct that we have internally is of service delivery excellence. This horizontal team works with all our industry verticals and project engagement. They run the voice of customers survey, and also continuously evaluate which are the top ideas that are coming up – in terms of innovativeness and impact, and what makes sense to surface to the leadership team. These are the ones that are also then encapsulated as ideas and innovations that we can take into our MBRs and QBRs. So, we have a formal construct that is goaled around all the best work within the organization. We have enabled this with appropriate technology support, such as Confluence and Seismic, so that knowledge dissemination is happening internally within the organization. These two are working well for us, and we’re able to capitalize on them and put good ideas in front of clients all the time.
AIM: Could you provide an example, while maintaining client confidentiality, where the ‘Voice of Customer’ framework was employed to balance innovation with immediate needs? We’d like insights into how these conversations typically start, your expectation-setting strategy, and the impact on project success.
Rajan Sethuraman: I can quote the example of a device manufacturer that we have been working with. They are one of the largest producers of printers, peripherals, and laptops. The problem statement that they originally came to us with was forecasting. They wanted to forecast the demand for the different parts that go into all their devices. That’s what we focused on initially – can we build good demand forecasting models based on past data to look at when the components are needed? This was focused on their maintenance services, as they usually have an AMC, especially with large B2B clients who are buying thousands of these devices, and the company is responsible for maintaining them as well. Interestingly, we were able to look at a movement from forecasting and forecasting-related metrics into a different kind of a project, where we could anticipate when the product might actually fail. Because the forecaster is looking at how to make sure that they have the part available when the device fails, so that they can take care of the client’s problem quickly. Whereas if you peel the onion, what you really want to do is to make sure that the device doesn’t even fail! Or even better, you’re able to go and make an intervention even before the device fails, because you know when it will fail. The project eventually moved in that direction. We started looking at whether there were enough data points available from all the sensors and instrumentation in the device, that will help us predict the possibility of when a particular component is going to fail. For instance, issues with a laptop’s motherboard often result from other factors, such as the improper functioning of the battery, fan, and heat dissipation unit. You have sensors and instrumentation on these devices that tell you if the fans start fluttering, then something is going to happen to the motherboard soon. You can build good ML models that look at data from all these sensors. We brought some good ideas to the table, suggesting that with this algorithmic approach, models can be built to predict upcoming issues. This allows for proactive intervention before a device fails.
We did a similar project in the automotive sector, collaborating with a major luxury German car manufacturer. In this case, the car was instrumented, and they approached us with a concern about unusually high warranty claims on specific parts. The challenge was to identify the cause, as these parts were expected to function for ten years but were failing after six or seven years. We were able to borrow ideas and constructs from our similar project and bring it into the context of this device manufacturer. So, it’s not always about looking for a new algorithm, but also going back and looking at what we have done and analyzing if it is relevant and applicable in a different context. It’s a combination of many things coming together that makes it possible.
AIM: Any concluding thoughts Rajan?
Rajan Sethuraman: The one thing that we keep hearing from clients is that even if you have the best analytics and decision-making frameworks and mechanisms, if people don’t use them, then all of it is futile. Today, analytics is not only about the latest innovations in technology, algorithms, or methods. There is a significant aspect of change management involved. There’s the saying, “Nothing succeeds like success”, which means that you need to show real progress on the metrics that matter. So, you cannot be innovating or ideating from an ivory tower, which is completely divorced from what is happening on the ground level. I think that balance is very important. However, given the kind of general excitement that is there today, in comparison to even say a year ago or 18 months back, it’s a great environment for figuring out how to leverage some of the available tools. It’s not that these technologies are completely new; large language models have been around for some time. Of course, they have taken different avatars, in the last several years, but what is changing now is the general excitement. So, clients can look at how to make use of some of these technologies to drive real action, from a P&L standpoint, for the metrics that matter.