“ If you are in a certain domain for a very long time you become more of a subject matter expert and depth-oriented person rather than becoming somebody who can think beyond the traditional way of approaching things. For AI & Analytics which is more of a horizontal function, it works well if you come with a cross-industry experience. If you see many of the successful leaders in the analytics space they don't come from a single domain and are generally able to set up teams with the curiosity to learn about the new domain. The cross-pollination of ideas across industries is what sets them up for success”

– Excerpt from the interview with Bharathram Ramakrishnan

Today is Episode 21 of the Interview series on Expert-Talks, with Thought Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Bharathram Ramakrishnan ( Bharath ), Global Head of Data Science and AI, at Novartis.  Before Novartis, Bharath was the Global Head of Data Science and Analytics at Dupont. He also held  Analytics leadership roles at Shell, TCS and Mu Sigma. Bharath holds a PhD in AI, an MBA in Systems and a Bachelor in Electronics Engineering.

We are listing below a few key points from the interview :

·         Bharath highlights the significance of understanding domain-specific challenges and being able to communicate effectively with business stakeholders. He emphasizes the need for collaboration within analytics teams and across departments to achieve success in delivering analytics solutions.

·         Key functional areas for business analytics in Pharma include operations, research, and distribution, each requiring high accuracy and reliability.

·         In the pharmaceutical industry, there's a push for faster delivery of impactful analytics, necessitating innovative approaches like synthetic data to bypass red tape.

·         Generative AI is predominantly used in research and development within Pharma, aiding in summarizing large volumes of documents for decision-making, while explainable AI remains crucial for ensuring safety, reliability, and compliance within the industry.

·         The interview touches upon the shift towards freelance and remote work arrangements in the industry, necessitating adaptability from both organizations and employees.

·         The top 3 areas an aspiring analytics professional needs to develop, are curiosity about domain challenges; effective communication with business stakeholders; and a collaborative work approach

 You can watch/listen to the interview on our Website, YouTube, Apple, Amazon Music and Spotify podcasts on the links given in the comments section below.  Please do share your comments and subscribe/follow us on @maavrus on LinkedIn, Facebook and Twitter.

The point is, how can we create an ancillary ecosystem based on Generative AI capabilities? So that is where the crux for people like me will be because I am not going to focus on creating a chat GPT backend engine. My focus would be on how I can use them for my business needs, right? How can I improve my product description based on an understanding of Google Search parameters, to help improve my organic search ranking, and hence improve my Marketing RoI – Excerpt from the interview with Bhargab Dutta

Today is Episode 20 of the Interview series on Expert-Talks, with Thought Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Bhargab Dutta, Chief Digital Officer at Century Plybvoards.  Before Century Plyboards, Bhargab was Director of Digital & Analytics CoE at Colgate Palmolive India Ltd. He has also held leadership roles in Digital & Analytics at Aditya Birla Group, General Mills and Honeywell. He is recognised as among the Top 10 Chief Digital Officers in India by CEO Insights Magazine and Top 100 AI leaders by 3AI.

We are listing below a few key points from the interview :

• Bhargab emphasizes the importance of data-driven decision-making, incremental growth, and cost-saving opportunities for organizations.
• Domain expertise plays a crucial role in articulating solutions and identifying the right data points.
• Analytics leaders should tailor their analytics journey & approach based on an organization's technology & Data maturity.
• Situational awareness is key in analytics and transformation journeys, allowing for immediate course correction and redefining solutions.
• New technologies, such as neurovision-based analytics , geospatial analytics and AI, are gaining ground in understanding end consumers and their buying patterns in the FMCG industry.
• Advice for Aspiring Data Scientists and Digital Professionals- Invest at least 5 years in one domain for depth of industry exposure. Focus on basics, especially statistics, before pursuing advanced topics like machine learning. Emphasize defining the problem statement before jumping into solution mode.

 You can watch/listen to the interview on our Website, YouTube, Apple, Amazon Music and Spotify podcasts on the links given in the comments section below.  Please do share your comments and subscribe/follow us on @maavrus on LinkedIn, Facebook and Twitter.

Attribution has been word talked about and made more complicated as well. Thanks to a lot of the analytics products and companies who have come through, I would say there is an inherent bias to let people not properly attribute. It is better to focus on the efficiency & effectiveness of each channel. By getting a very strong UTM framework implemented at each channel, you will be able to tell for eg that 50% of my traffic comes from organic, 30% comes from Google, and Facebook, 10-20% comes from CRM, and thereby help maximise the effectiveness of each channel for eg in performance marketing can I reduce that bidding on my keyword so that I can let the organic traffic flow?

– Excerpt from the interview with Saurabh Agrawal.

Today is Episode 19 of the Interview series on Expert-Talks, with Thought Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Saurabh Agrawal, Founder & CEO of DAIOM ( Data and AI in OmniChannel ).  Before DAIOM, Saurabh was SVP of Analytics and Growth Marketing at LENSKART. He has also held leadership roles in Digital, AI and Analytics at MothersonSUMI,  Tata Insights & Quants and American Express. Saurabh speaks frequently at industry forums on leveraging marketing analytics to enable profitable business growth.

We are listing below a few key points from the interview :

• The key areas where analytics plays a role in D2C brands are growth marketing, customer experience, and financial profitability. These areas help drive business growth and increase customer engagement.

• D2C brands are expanding into physical stores to bridge the gap between online and offline customer experiences. This allows customers to overcome challenges like size and product quality, and helps build brand identity and trust.

• Digital native brands have an advantage in terms of tech agility and nimbleness compared to traditional brick and mortar retailers. Process digitization and consistent store experiences are essential in ensuring a seamless customer journey.

• Data availability and quality are challenges in the omnichannel world. Customer identity management and unifying data from different tech systems are crucial for effective analytics and personalized customer experiences.

• Influencer marketing can be a cost-effective way for brands to reach their target audience. It is important to choose the right influencers based on their audience, engagement, and relevance to the brand, and to measure the impact of influencer campaigns using attribution and analytics.

• Balancing digital marketing spend across different channels is essential for reaching the right audience. While traditional media can still be effective for brand recall, digital channels offer more targeted and measurable results. Consistent and measurable efforts in organic and influencer marketing can lead to a higher ROI.

 You can watch/listen to the interview on our Website, YouTube, Apple, Amazon Music and Spotify podcasts on the links below.  Please do share your comments and subscribe/follow us on @maavrus on LinkedIn, Facebook and Twitter.

For a long time in the industry, I think we made the mistake of going behind people who were experts in let's say certain set of algorithms; who understood the math very well, but they were not able to convert that mathematical problem or match the mathematical problem to a business context. So without context, they are just algorithms and numbers and libraries and codes which may or may not solve the actual business problem– Excerpt from the interview with Mathangi Sri Ramachandran. 

Today is Episode 18 of the Interview series on Expert-Talks, with Thought Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Mathangi Sri Ramachandran, Chief Data Officer at YUBI.  Before YUBI, Mathangi headed Data Science functions at GoFood ( part of GoJek ) and PhonePe.  She has near 20 years of experience in Data science and analytics, 100+ patents to her name, is an author of 2 books and is among the Top 50 Influential AI leaders in India.  

We are listing below a few key points from the interview :

• Data is an organisation’s asset. So if data access gets restricted to only the data science and analytics team, then democratization and large-scale data-driven decision-making will not happen. So depending on how friendly the end users are with technology, you build an application layer and democratize these components so that each can explore their insights.

• In a simplified manner, Mathangi brackets analytics initiatives into  1. Human-led and Machine-assisted, where you provide self-serve models and data access in intelligent formats to businesses,  for them to make decisions and act quickly and 2. Machine-Led and Human-governed, where the analytics team is building transformative break-through models, with high and long-term business impact.

• For the Data Science team to create a meaningful impact in an organisation, the pre-requisites are the Quantum of available data, Business problems with constraints and an Organisation culture that encourages experimentation.

• An organisation that is open to experimentation & failure, will encourage teams to explore probabilistic scenarios and take big bets for non-linear impact. The absence of this can lead teams to be more deterministic in their approach, which by nature will be incremental in impact.

 You can watch/listen to the interview on our Website, YouTube, Apple, Amazon Music and Spotify podcasts on the links below.  Please do share your comments and subscribe/follow us on @maavrus on LinkedIn, Facebook and Twitter.

For Analytics & AI professionals in GCCs to be successful, developing domain knowledge and context is very important. I really think that it has to come from within. I was always intrigued by retail as a domain and curious about how things operate in the business. And I was always trying to make sense of what the numbers were telling me and What does it mean?. So I was always trying to put myself in the shoes of the business. And that's something that I enjoyed.  – Excerpt from the interview with Vinodh Ramachandran

Today is Episode 17 of the Interview series on Expert-Talks, with Thought Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Vinodh Ramachandran, Head – Data Science & Analytics, Neiman Marcus Group.  Prior to Neiman Marcus, Vinodh was Site Leader and DVP at Saks Off 5th, where he led all their business functions including analytics. He has also held Analytics leadership roles at Lowe’s , Target and Genpact. Vinodh frequently shares his thoughts at Industry forums.

We are sure you will benefit greatly from listening to his perspectives. A few key points from the interview :

• Vinodh articulated the key factors that help make analytics initiatives successful.  1. A well-defined problem statement 2.  Alignment with the organisation's goals  3. Sponsorship from Top leadership. 4. Level of data maturity as measured by single source of truth and 5. The ability to explain the solution & insights to the business in an understandable and practical manner.

• To develop domain knowledge and context, reading financial performance reports about the company and competition, and understanding the company’s organisational structure & processes from the company’s intranet pages, helps to get an overall perspective of business.

• Exploring the data structure in a warehouse to understand product hierarchies, exploratory data analysis to understand customer behaviour and breadth of offerings, and then validating them in discussions with business stakeholders also helps in further building contextual knowledge.

• Usage of external & outside-in data, apart from helping an analytics professional build trust and connect with the business, also helps develop strong hypotheses when looking for insights; for eg number & density of pawn shops in a retailer’s catchment, could have a correlation to a electronic / luxury retailer’s store shrinkage.

• Generative AI is being used by many retailers, beyond generating content for customer engagement. It is also finding usages in other areas like customer service, where it is being used to summarise feedback from thousands of customer reviews to give a quick 80-100 synopsis to prospective shoppers.

For continued success, it is necessary that Data Scientists are perceived as business function/process experts by the Business stakeholders. Apart from spending time on the operations floor,  signing up for industry certification courses can help Data Scientists, build good credibility with business  - Excerpt from the interview with Neil Srinivasan.

Today is Episode 16 of the Interview series  “Expert-Talks @ MAAVRUS” with Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Neil Srinivasan, Founder & Managing Principal of Canopus Business Management Group. Prior to starting Canopus, Neil was SVP – Customer Experience and Service Excellence at HSBC. He also held Business Excellence Leadership roles at Bank of America and Stanchart.  Neil is a Six Sigma Master Black Belt and author of 3 books.

We are sharing a few key points from the interview :

• Banks traditionally are service-oriented, for which they need to continually drive efficiency into operations. To deliver this change one needs to be able to look at the end-to-end process and make fact & data-based decisions/solutions.  A strong background in Six Sigma process excellence makes it relatively easy to adapt to different industry domains.

• In B2B the customer data exhaust follows a 1:10:100 during the “Early stage”: “Pre-order stage”: “Post-order account mining stage”. Conversely, in B2C the consumer data exhaust is significant before the order and comes down during product/service consumption.

• Focus AI and Model efforts,  on customer or business segments, where there is a possibility for maximum impact and minimum time required for the model to be implemented, to start showing results. Neil calls this the “local-local” approach as opposed to a large global end-to-end project.

• Increasingly customers are buying experience and not the product alone, so manufacturing organisations are required to incorporate the variabilities of customer and employee behaviour, in the way they design their product and consumption experience.

 You can watch/listen to the interview on our website, YouTube, Apple, Amazon Music and Spotify podcasts on the links below.  Please do share your comments and subscribe/follow us on @maavrus on LinkedIn, Facebook and Twitter.

For any Business Transformation Leader to deliver on expectations, it is necessary that the leader is (i). In the transformation (ii)  Has a good team on board, and (iii). Is able to align the pace of transformation to the organisation’s pace. - Excerpt from the interview with Piyush Chowhan.

Today is Episode 15 of the Interview series  “Expert-Talks @ MAAVRUS” with Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Piyush Chowhan, Chief Information Officer, Panda Retail Company, Saudi Arabia.  Prior to Panda Retail, Piyush was the Group CIO  Lulu Group, UAE. He has also held Technology and Transformation leadership roles at Arvind Lifestyle Brands, Walmartlabs and TESCO.  Piysuh speaks frequently at Industry forums about the technology and customer experience advancements in Retail.

We are sharing a few key points from the interview :

• One key business expectation of an Omni-retailer is its ability to serve customers seamlessly from its stores/warehouses.  So a Single View of Inventory including the systems, processes and controls to ensure near-accurate inventory data across locations, and an optimised last-mile delivery model, are the first and most important steps in the transition to being an Omni-retailer.

• The important thing is having a 360 view of the customer. For a great Omni Customer experience, map the customer journey from an overall interaction perspective starting from pre-commerce to post-commerce, and then look at reducing the friction at each handover. One of the biggest challenges today is that most retailers are focused on improving only the commerce part of the experience, which leads to an overall disjointed experience.

• Understanding Cultural nuances across geographies is extremely important for a successful transformation project. Depending on the geography, a leader needs to balance the level of prescription and empowerment.  The same proportion will not work across different geographies.

• AI and ML programs & projects have to be embedded as part of larger digital transformation initiatives, for businesses to see the impact in terms of business outcomes. That is why ideally Data & analytics should be part of the Chief Digital Officer or Chief Transformation Officer’s organisation so that there is an aligned enterprise transformation journey.

You can watch/listen to the interview on our website, youtube, apple, amazon music and Spotify podcasts on the links below.  Please do share your comments and subscribe/follow us on @maavrus on LinkedIn, facebook and Twitter.

Youtube Video link. https://youtu.be/KeY98y7NylQ

People still trust people. While for simpler processes, we may trust machines because of their consistency & speed, when it comes to complex decisions or where the stakes are higher, we will continue to depend on people.  So for the foreseeable future soft skills will continue to be important in terms of human interaction, to get businesses to invest in areas that lead to higher customer experience & satisfaction. - Excerpt from the interview with Shailesh Jain

Today is Episode 14 of the Interview series on Expert-Talks, with Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Shailesh Jain, Group Head – Analytics & Insights, Landmark Group, Dubai.  Prior to Landmark Group, Shailesh was Senior Vice President & Head of Analytics ( Decision Management ) at Citibank India. He has also held Analytics leadership roles at KPMG Advisory and DunnHumby. Shailesh frequently shares his thoughts at Industry forums.

We are sure you will benefit greatly from listening to his perspectives. A few key points from the interview :

• Shailesh spoke about the top 3 building blocks he focuses on,  to build great analytics capability  & business impact. First, deep business partnering with the business & functional leaders to understand their needs & vision better; Second, his analytics team spends quality learning time in the process be it the retail stores/distribution centres/merchandising process etc; and third role-rotation across various functions for each person to understand the interlinkages and the big picture better.

• Projects are ideally planned with short-term objectives to create near-term business impact & credibility,  and with long-term objectives which align with the strategic transformational vision for that business.

• Getting outside in perspective to complement internal performance has always been a key business expectation, and was mostly collected manually through surveys, research etc. This has now become easier thanks to customer’s digital footprint across social media, search, website journeys etc.

• In future, any processes where decisions are based on business rules or discriminative insights will see faster AI adoption,  while complex decision making which requires a leap of faith / probabilistic calls, will still be taken by humans.

• When it comes to generative AI,  organisations for now will use it for in-house processes within their private cloud. For any external & customer-facing use cases, it will continue to be with a human-in-the-loop

• A Stanford University research validates that Deep Learning models with Explainable AI outperformed other deep learning models.  This is because, rarely does a model work in isolation – typically the output of one model is fed into another model, and so understanding the variables & weights of each model is important.

 You can watch/listen to the interview on our website, youtube, apple, amazon music and Spotify podcasts on the links below.  Please do share your comments and subscribe/follow us on @maavrus on LinkedIn, facebook and Twitter.

Youtube Video link. https://youtu.be/zOfYCdGNYOc

The key to being a good data scientist is to behave like a child – stay curious and willing to learn. Use this approach to continuously improve in the four areas (i) good domain knowledge (ii) foundation in maths & analytics (iii) comfort with tech /programming (iv) business empathetic communication skills, and success is guaranteed - Excerpt from the Expert Talks interview with Dr. Manish Gupta.

Today is Episode 13 of the Interview series on Expert-Talks @ MAAVRUS, with Leaders in the Analytics, AI and Transformation space. For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Dr Manish Gupta, Global Data Science leader at Microsoft. Prior to Microsoft, he was Vice President and CoE Head – Machine Learning & Data Science Research at American Express. He has also held Analytics leadership roles at InfoEdge and Citibank. He is a PhD in Mathematics from IIT Delhi.

We are sure you will learn greatly by listening to Dr. Manish Gupta. A few key points from the interview :

The basic difference in analytics solutions between digital and physical businesses is the volume and velocity of data. Digital companies have a higher appreciation of data and have a data-first mindset since it is core to their survival. They need to leverage the data for impacting customer experience and retaining them.

For traditional companies data and AI can provide a competitive edge. The early adopters, build a data culture quickly so that they can stay ahead.

For building a culture of data and AI, business leaders need to have trust in analytics as an engine for growth. And it has to be reciprocated by the analytics team by developing impactful solutions & articulating the ROI both for business and consumer benefit. It's a virtuous cycle for building data culture across the organisation, thereby motivating the analytics & business user teams.

Generative AI will democratise the usage of data sciences. Even software developers, can plug-in some of the APIs , use appropriate prompt engineering and do wonders for larger community benefit. This is absolutely an inflexion point in the adoption of data science-enabled value creation

There is a need to bridge that gap, so that businesses can appreciate the technology and technology can benefit the business. This is where systems like generative AI can be very helpful. ChatGPT and Microsoft Co-pilot are enabling business users to engage in conversational tones to get access to decision-enabling insights.

You can watch/listen to the interview on our website, youtube, apple, amazon music and Spotify podcasts on the links below. Please do share your comments and subscribe/follow us on @maavrus on LinkedIn, facebook and Twitter.

Cross-industry knowledge/exposure could be a great asset for Analytics & Transformation leaders. Your customers are getting influenced not just by particular competition trends in your industry, but also through touchpoints from other industries.  So let's say you are in a B2B company and are interacting with a client’s Procurement manager. That person in his personal life is shopping online, and is exposed to single-click ordering, real-time delivery status updates and feedback collection etc. It is very normal that the same person now wears the hat of a B2B customer and then expects a similar experience from you as a  supplier. - Excerpt from the Expert Talks interview with V Ganapathy

Today is Episode 12 of the Interview series on Expert-Talks, with Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with V Ganapathy, Vice President and Head of Global Advanced Analytics CoE at Holcim. Prior to Holcim, Ganapathy was Senior Director and Head of Business & Enterprise Analytics at Philips. Previous to that, Ganapathy has worked at  Dell, AOL, Ford and MRF. He is a thought leader who frequently speaks at AL & Analytics Forums.

We are sure you will enjoy listening to Ganapathy and his perspectives. A few key points from the interview :

• For analytics to be impactful, there are a few factors which are important. Firstly Culture, secondly AI & Analytics ecosystem Capabilities and thirdly Domain knowledge & Business connect.

• Culture should encourage continuous learning of new research and best practices; an experimentation approach to continuously test & learn and fail fast/cheap; and a curious mindset which is scanning the status quo to identify new opportunities.

• AI and Analytics ecosystem capability that covers agile ways of working, aligning and executing to business priorities; secondly data assets that are clean and harmonised to identified initiatives; platforms that enable rapid experimentation and flexibility in seamless scaling; embedded analytics capability in each business unit; and the governance to track the performance linkages based on the above constituents.

• One of the decision points for AI leaders is about buy vs build. What are those industry-agnostic models that can be customised and deployed quickly, and what needs to be built grounds-up? A great approach is to do the feature engineering in discussion with the business. This is where domain knowledge and the ability to engage business stakeholders become extremely important

• Both quick turnaround, as well as strategic projects, are important. Quick wins help build business confidence in the AI teams, which is important to get investment to support the long terms strategic ambitions of the AI team for the business. Quick wins also play a role in bolstering the confidence of the Analytics teams, which is important for them to pursue a mindset of curiosity and experimentation.

• When prioritising investments in data ecosystems & AI projects, using a 2 X 2 of Data Intensity Vs Enterprise Value opportunity of that function /process, is a great approach. Ideally where both data intensity and enterprise value is high, is a good sweet spot.

• Cross-domain projects typically provide great monetary value as well as strategic leverage. A customer journey value stream, is a great way to identify such projects. Having said that such projects could be complicated given the need to establish multiple stakeholder buy-ins and possibly differing levels of data availability & maturity across the functions.

“Do away with the Superwoman syndrome because what happens is in our own heads we keep thinking, oh, I am expected to be a superwoman. I'm expected to do well on the job and I'm expected to do well at home. I would say don't do that to yourself. Do a fair share of work at home. Do a fair share of work at the office. There is only so much that you will be able to do, but whatever you do, do a great job of it. And don't shy away asking for help.”  - Mamta Rajnayak’s advice to women professionals. Excerpt from the Expert Talks @MAAVRUS interview with her.

Today is Episode 11 of the Interview series on Expert-Talks, with Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Mamta Rajnayak VP - Head of AI-ML Products & Platforms@AI Labs , American Express. Prior to American Express, Mamta was managing Director and Retail Analytics Leader at Accenture Global AI Hub. Prior to Accenture, Mamta worked at Adobe, Evalueserve and ICICI Bank. Mamta holds 5 patents and is a frequent speaker at Analytics & AI industry forums.

Listeners will find Mamta’s perspectives very insightful and enriching. We are listing below, a few key points from the interview :

• Mamta mentions that as a Senior Analytics Leader, apart from coming up with the right framework and bringing the right team on board - who will be able to deliver the work well,  one should also be able to understand stakeholder psychology and where they are coming from, address their concerns in an amicable manner, so as to get solutions implemented. Having said that, she also believes that solid technical skills are a must-have because that's why one is there in the first place. It is also important to understand the business problem at hand and frame it in such a way that the data would be able to solve it.

• For a project to be successful, it is important to develop a stakeholder power map. Understand who are the decision makers and what are their priorities. For eg, a Fraud Prevention Leader or Risk Assessment Leader would be more mindful of accuracy levels and may not be time-pressured, whereas a marketing leader may be okay with lesser levels of accuracy, but may want quick turnaround iterative inputs. Getting this clarity during the project planning stage can significantly enhance the chances of success.

• Any progressive organization should have an external perspective at the centre of its decision-making. You have to understand how the industry is progressing, where the movement is, and if,  growth in a certain technology is going to benefit you or not. Today there are multiple external data sources; you can use data from Google for eg search history of a person,  location data mobility data, credit ratings/scores from 3rd party data providers, and census/survey data.  You could also try collaborations with other different businesses.

• Low Code/no-code platforms will help analysts and data scientists be more efficient and free them of tasks that are repetitive and which can be automated. This will mean that the analytics teams will have a lot more time to engage with and understand the business, think more creatively and innovatively, and focus on implementing the solution practically in the business.  However, Data scientists and analysts will always need to have critical thinking skills and an understanding of data sciences from first principles.

“People run organisations, technology doesn’t. So it is important to explain the WHY behind the recommendations of an AI model to business leaders and users. The whole field of Explainable AI is becoming increasingly important; it started with SHAP and LIME which help go deeper behind the features in AI models. Currently “counterfactual explanations “ is another emerging area in the explainability space.” - Excerpt from the Expert Talks interview with Anirban Nandi.

Today is Episode 10 of the Interview series on Expert-Talks, with Leaders in the Analytics, AI and Transformation space. For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Anirban Nandi, Vice President – AI Products & Business Analytics at Rakuten. Prior to Rakuten, Anirban has had analytics programs at Landmark group and Target. Anirban is a frequent speaker and Analytics and AI industry forums and is also an extensive blogger on AI topics.We are sure listeners will enjoy listening to Anirban’s perspectives over the next 40 minutes. We are listing below, a few key points from the interview :

1. Anirban emphasizes the key building blocks required for a successful AI and Analytics program in any organisation. Firstly, there has to be organisational and business alignment and belief system in the AI and Analytics program, secondly, there has to be a vision and data strategy with well-articulated KPIs and goals, thirdly the company needs to invest in collecting and storing quality data with the appropriate data latency and fourth it needs to build the right team, platform and tech capabilities to convert the data into actionable insights and models. Most importantly the AI and Analytics teams need to invest in developing a business context which is relevant to its specific industry and market situation.

2. A combination of quick turnaround analytics and deep breakthrough AI models may be required to create both immediate as well as long-term value for organisations. For companies which are at initial levels of AI and Analytics adoption/maturity, the proportion may be skewed more towards quick turnaround analytics & insights. However as an organisation matures, it can automate more and more insights and focus on building transformative AI models and integrate them into its delivery/platform features.

3. It is always a good idea to get outside in perspective and data and augment it with internal data, during an organisation’s digital transformation journey. While the products that customers buy are different in different industry segments and have their own nuances, some of the basics like why and when customers buy can be applied from first principles across industries. Businesses can also leverage external anonymised data from external providers, data from GEO SDKs and in some cases avail the services of consulting firms to build more outside-in perspectives.

4. Generative AI is an augmentation or extension to the existing field of AI, and will soon large disruptive adoption by businesses, in many areas like EdTech, Contact support etc. As it begins to get wider traction, Prompt Engineering will become an increasingly important skill. Business Transformation Professionals who have a good understanding of business constraints/ outcomes, and the possibilities of Tech / AI, are most likely best positioned to leverage their experience for better prompt effectiveness.

5. Future of AI and Future of Work will definitely influence each other. AI will not take away your job, but a person who uses AI effectively could take away your job. With increasing adoption of generative AI, the nature and expectations from each job/role will change and professionals will need to stay abreast, upskill themselves and adapt to changing environments.

Today is Episode 9 of the Interview series on Expert-Talks, with Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Nitin Srivastava, Data and Analytics Leader at Advanced Auto Parts.  Prior to Advanced Auto Parts, Nitin has had a wonderful career at UnitedHealth, Wells Fargo and Mahindra Satyam. 

We are sure listeners will greatly learn from Nitin’s practical examples of business value creation through the application of analytics. We are listing below, a few key points from the interview :

• Nitin emphasizes the importance of understanding the business problem clearly before working on any analytics projects. Nitin has worked with various companies operating in different industries, and he explains that each industry has its nuances, which is essential to understand to build capabilities effectively. He highlights the need to be flexible and adaptable to the culture and level of analytics maturity in each company to be successful as an analytics leader.
• With regard to the future of work and how AI can play a role in it, he believes that AI will play a major role in connecting developers with companies and clients. AI-powered suggestions could connect the right people across the globe. Nitin sees AI playing a significant role in building industry-specific domain expertise, such as a large language model in the retail industry that is specific to the auto parts sub-domain. Nitin believes that AI will play a major role in every walk of life and not just for data and analytics folks. He notes that embedded AI will be playing a vital role in every person's life.
• He acknowledges concerns about AI taking away jobs but points out that new jobs have been created in the past, such as during the introduction of the steam engine and calculators. Nitin believes that AI will create new job opportunities, and people need to adapt to these changes by acquiring new skills. Prompt engineering is a new skill that Nitin believes will be crucial in the future, as it is necessary to interact effectively with AI. He explains how Microsoft 365's Copilot can create a PowerPoint presentation by extracting information from Word documents and voice prompts. Another example is GitHub Copilot and its potential to increase productivity and reduce the time to market.
• To have a successful career in AI, analytics, and transformation, budding data analysts will need to, firstly, have a passion for the job, secondly, they will need to develop business acumen to understand the business side of it and thirdly, they need to have a good understanding of algorithms and where to use them, which includes supervised and unsupervised learning techniques

You can watch/listen to the interview on our website, youtube, apple, amazon music and Spotify podcasts on the links below.  Please do share your comments and subscribe/follow us on @maavrus.com on LinkedIn, facebook and Twitter.

Today is Episode 8 of the Interview series on Expert-talks @MAAVRUS,  with Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Vijoe Mathew, Senior Global Director - Analytics at AB InBev.  Prior to AB InBev, Vijoe was Advanced Analytics solutions leader at Honeywell. He has also worked across all areas of retail analytics at TESCO.  Vijoe is an Engineer, Masters in Industrial Management and Six Sigma Black Belt. 

Listed below are, some key points from the interview :

• Vijoe articulates a few success factors for analytics. Firstly, analytics must serve a business purpose and create value for the company. Secondly, it is important to partner with subject matter experts to understand the business problem statement. Thirdly, it is crucial to have a good understanding of data maturity and invest in building the right skill set to build analytics as products rather than just services.

• Business should receive the credit for the solution, rather than the analytics teams taking the laurels. He believes that data science is domain agnostic, but business collaboration and domain inputs are necessary to make it a success. He believes that analytics teams should work closely with business owners who share the same passion to make solutions successful.

• Analytics can be used for both tactical and strategic purposes. He notes that today, everyone in a business is measured on value creation, which has led to a shift from efficiency-based KPIs to focusing on data-driven value creation. Vijoe explains that creating solutions that address today's problems and creating a pipeline for future solutions is crucial.

• He emphasizes that the credibility of the value created through analytics projects is necessary to attract business stakeholders and sponsors, and to improve the acceptance and reach of analytics in a company. Attributing success can be a challenge, especially in top-line growth-related initiatives, since at a particular point in time there could be many projects with overlapping benefits.  There is a need for a framework and standard processes to ensure that solution benefits go through a test versus control methodology and are approved by the relevant teams.

• To understand customer behaviour in the CPG industry, companies need to partner with the right data providers to gain insights into the market, customer preferences, and competitors. Vijoe also discusses the importance of an outside-in perspective from data, mentioning that partnering with the right retailers and Point-of-consumption chains can provide valuable data to create tailored offerings.

• For aspiring data scientists, he stresses the importance of continuous learning, as technology is constantly evolving. He mentions the need for leadership skills and the ability to work with the business to understand and solve problems. Vijoe encourages a transformative mindset and problem-solving skills, as these are essential to stay relevant in the field. He concludes by stating that technologies will come and go, but great problem solvers will always be in demand.

Today is Episode 7 of the Interview series on Expert-talks @MAAVRUS, with Leaders in the Analytics, AI and Transformation space. For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Utpal Chakrborty, Chief Digital Officer at Allied Digital. Prior to Allied Digital, Utpal was Head of AI at YES Bank. He has also worked as principal architect in Capgemini, IBM and L&T Infotech. He is an eminent speaker, researcher and author on topics ranging from Artificial Intelligence, Quantum Computing, Web3 and Metaverse, Blockchain and IoT. In 2022, he was conferred with the award of Global AI Ambassador. I am listing below, some key points from the interview : 1. Utpal is interested in various technologies because of their unique features. Quantum computing can be linked to artificial intelligence through quantum machine learning, which boosts speed and agility. Web 3 and Metaverse are interlinked with artificial intelligence and blockchain. 2. Multiple factors contribute to a country's AI ecosystem, and India has evolved significantly in this regard over the past few years. While the US and Canada lead in AI research, India has made significant progress in implementation, particularly in its startup ecosystem and response to digital payments and the COVID pandemic. 3. Allied Digital has been involved in Smart City projects in India for a long time, and they have gained experience in implementing various use cases. They have evolved from infra-level implementations to more intelligent systems that incorporate computer vision and advanced analytics. The aim of a Smart City is to build a digital infrastructure that provides better customer experiences for citizens and makes government assets more productive. 4. The key components of a Smart City project, include smart lights, garbage management, traffic management systems, and sensors. The data collected by these systems need to be brought together in a single repository using ETL processes and then transformed, correlated, and filtered to build data marts. Machine learning models, AI models, and analytics can be applied to derive insights and build predictive models that can be used by the command and control Centre to take immediate action or by citizens to access intelligent services through their apps. 5. Domain support, situational awareness along with support from top management is important for the success of Analytics & AI projects. For achieving scale data engineering is a critical aspect. Given the all-pervasive requirement of analytics and AI in driving transformation, it requires higher-level leadership to bring multiple teams together towards a common goal. 6. Utpal sees AI and the future of work converging, making it easier for people to work seamlessly and collaboratively. Large language models like ChatGPT will become more powerful and will be applied to other mediums such as images, videos, and audio. He believes that outcome-oriented work will replace conventional nine-to-six jobs and that agility will increase as AI is applied to tasks that currently take longer.

Today is Episode 6 of the Interview series on Expert-talks @MAAVRUS, with Leaders in the Analytics, AI and Transformation space. For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Gary Cokins, Founder & CEO of Analytics-based Performance Management LLC, a company that provides financial insights and analytics. Gary has previously worked with KPMG, Deloitte, EDS and SAS Institute. Early on in his career, Gary worked with Dr Robert Kaplan and Dr David Norton, the creators of the balanced scorecard. Gary completed his BSc degree in industrial engineering and operations research from Cornell University, and MBA from Northwestern University’s Kellogg School of Management. He has also authored many books on activity-based costing, and co-authored a book “Predictive Business Analytics” I am listing below, some key points from the interview : Analytics in enterprise and corporate performance management is gaining traction, for a few reasons like executives frustration with strategy failure, increasing scrutiny of their performance, need for rapid decision making & the accompanying risks, and mistrust of management accounting systems. For Business Unit leaders, overhead cost allocations not being transparent or being too simplistic is a challenge, and organisations are adopting activity based costing to be able to better relate investment/ costs to outcome. Activity based costing is also important to understand Customer Lifetime Value (CLTV), which is a predicted measure of a customer’s profit contribution over the estaimted period, that she / he is likely to be shopping with the brand. Its classical definition is, discounted cash flow net present value prediction for a customer. After accounting for individual expenses like distribution cost, channel expenses, marketing spend, discounts , cost-to-serve etc, organisations most often realise that their largest customer(s) by sales, may not be the most profitable – because they could adding cost elements like non-standard asks, changes in delivery schedules, higher frequency of product returns etc. Hence Business and Marketing teams need to better understand which customer segments are most attractive, whom they need to acquire, retain, win-back and grow. However, most organisations prefer to stay with the out-of-date annual budgeting process, since doing activity based costing requires investment in process, data and time. Also looking beyond the enterprise, organisations need to adopt a collaborative approach with suppliers ( as opposed to being adversarial ). That is because in the changing world, supply chains are competing against other supply chains for end consumer share of wallet. You have to embed analytics in each of the methods / processes, that can help deliver effective measurable and consistent performance. Analytics and Change are like gears and machine, and they have mesh seamlessly. Good Data scientists have to “Search for Surprises”, and build a strong data story, to challenge ingrained ways of leadership working. So it is combination of business understanding, collaborative hypothesis building, exploratory approach and ability to convince with great narration. If there is discomfort with current situation (D), a well-articulated vision of the end state /future (V), and if the First steps to get there are practical (F), then their combined might, will help to overcome Resistance to change (R). So D * V * F should be greater than R. Change leaders will do well to build and validate narratives along the above lines. Explained differently, Transformation and Analytics teams should answer not only the “WHAT”, but also the “SO WHAT” and “THEN WHAT” The future of AI will disrupt the way work is delivered today, and so anything that is repetitive and can be automated will be done by machines, and humans to need focus and build higher order business decision making and people skills.

Today is Episode 5 of Interview series on Expert-talks @MAAVRUS, with Leaders in the Analytics, AI and Transformation space. For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Rob Hand, Founder & CEO of Hand Promotion Management, a company that provides advisory services and transformation consultation to Global CPG and Retail Organisations, in the areas of Trade Promotion, Retail Execution and Revenue Growth Management. Previously Rob has worked with Capgemini, SAP, Oracle and Media Net. He is also author of the book “Invisible Economy of Consumer Engagement” We are sure we can learn a great deal from Rob’s insights based on his extensive domain experience. 

A few key points from the interview : 

Around 19-20% of CPG Sales turnover is spent on Trade Promotion, making it the 2nd largest cost line for CPG & FMCG companies, and hence there is a need for a lot more linkage between Trade Promotions and the end Consumer engagement and outcomes. This requires a good understanding of Trade Promotion Planning, Execution and Revenue Growth Management processes. With Covid lockdown strangling businesses, Rob decided to write the book “Invisible Economy of Consumer Engagement” to explain the current on-ground situation and his thoughts for the future. 

Two-thirds of trade promotions do not deliver on business expectations and are categorised as failures – a key reason being the availability of trade promotion performance data. Even existing data with CPG companies are lacking in accuracy, granularity and scope. To be able to apply sophisticated AI/ML engines, will require data with highly concentrated hierarchical detail. 

Historically, Grocery retailers have looked at Trade Promotion spending by CPG companies as a way to boost their thin profitability margins. So while retail data which is rich and timely in customer shopping behaviour and experience, can help the overall product value chain, there has been a reluctance to transparently share data. However, given the rise of e-commerce during the pandemic and the fact that it has remained at record levels, the brick and mortar retailers know they have to fight back smarter, and that means sharing intelligence with CPG companies to execute productive promotions. 

A traditionally painful area of trade and channel promotion is the full reconciliation and settlement of deductions taken by the retailers and distributors where, due to the lack of visibility, the supplier has difficulty identifying the source or reason for the deductions. For most CPG companies the deduction could range between 20 to 40K deductions per month, and write-offs in the range of approx. 3% of that value. Lack of historical data granularity is one of the reasons, why auto-reconciliation continues to be a challenge even with sophisticated trade promotion solution vendors. 

Another reason for trade promotion failure is the lack of alignment between the trade calendar and the corporate marketing event calendars – coupons, e-commerce events, advertising, and so on. With most CPG companies also investing in their own digital commerce channels, a portion of their consumer marketing spend is used up in immediate gratification through coupons etc, putting pressure on their brand-building investments. Most CPG companies have now adopted Revenue Growth Management roles to be better prepared for the Omni-world. His book “Invisible Economy of Consumer Engagement” explains, 4 stages to the level of consumer engagement. The final stage “Engaged” is almost utopian – and means no failed promotions, which means accurate, granular and trusted data, NO out-of-stock conditions, 100% alignment between trade promotions, corporate marketing and ecommerce promotions, and it means 100% execution of the integrated business plan every day of the year. Most companies are though in the first or second stage of the ladder.

Today is Episode 4 of Interview series on Expert-talks @MAAVRUS,  with Leaders in the Analytics, AI and Transformation space.  For this episode, our CEO Mahadevann Iyerr (Mahaa) is in conversation with Rahul Pednekar, Vice President & Head – Advanced Data Analytics, Actuarial and Data Insights at Swiss Re. Prior to Swiss Re, Rahul has had a wonderful career at Vodafone, JP Morgan Chase and Infosys.  He also frequently participates in AI & Analytics hackathons, and speaks at Industry conferences on the Applications & Future of AI.

We are sure listeners will greatly benefit from Rahul’s depth of knowledge, and his thoughts on achieving analytics success, by using a practical project lifecycle.  We are listing below, a few key points from the interview :

• According to Rahul, Data Science is an intersection of Programming, Maths & Statistics and Domain understanding, and to be successful one needs to have a well-rounded ability across all these aspects.
• Data scientists cannot learn in isolation. They should work and collaborate with other data scientists to learn more.  Open hackathon platforms are a great place to ask, share and learn. Also when participating successfully in hackathons, it is always a good idea to create a digital footprint of your approach through blogs etc. Apart from getting a sense of validation, it could also inspire fellow data science professionals to try & learn.
• To build domain knowledge, one should stay updated by reading industry articles and magazines and by interacting with business domain practitioners and learning from their viewpoints.
• From a data scientist’s perspective, it is also necessary to spend time learning and building models on their own, so as to appreciate any industry-specific nuances,  for eg for actuarial modelling R is better suited than Python.
• When embarking on an ML transformation project, it is always good to build an initial prototype to check for business alignment on the controllable input factors, and possible outcomes, as well as get a sense of the constraints from a data availability & quality perspective.
• External data is very important and helps build a market perspective for the business. However in many cases, real-world data may not be available, so one has to look at trusted sources for forecasted data about possible input factors. For eg when it comes to global economic or financial data one could explore sources like Moody’s, Bloomberg etc.
• Data scientists will need to be able to build models, where they can help business understand in a simple manner, why a model is predicting in a particular way, and also justify how the outcomes predicted by the model is in alignment with stakeholder expectations/business case.
• For analytics / ML projects to be successful, Data scientists need to have a larger understanding of data engineering, data ops, model building and endpoint creation through ML ops. Otherwise, the model is likely to stay just on their Jupyter Notebook, and not get operationalised.
• Organisations are still trying to understand the possibilities and limitations of OpenAI and deep learning platforms. While there are a lot of innovation and business transformation possibilities with Open AI, there are still concerns from an ethical viewpoint, unintended bias, computing capability and investments.
• Directionally the future of AI will be a collaboration between the ingenuity of humans and the computing scale of machines.
• Young aspiring data scientists will need to stay curious and hungry for continuous learning, build a strong foundation in statistics, and collaborate extensively with data scientist communities, both within and outside the company.

Today is Episode 3 of Interview series on Expert-talks @MAAVRUS,  with Leaders in the Analytics, AI and Transformation space.  For this episode, I am in conversation with Saswata Kar, Senior Director and Global head of Data, Analytics and Data Sciences at Philips Global Business Services.  Prior to Philips, Saswata has had an illustrious career at Capital One, HSBC and GE Capital.  He is also a Forum Member of Nasscom CoE for IoT and AI.

It was wonderful to speak with Saswata and understand how he uses  his background in economics, statistics and corporate finance, to provide impact creating insights & analytics solutions. It was great to learn from his business empathetic practical approach. Am listing below, some key learning from the interview.

• He feels that his majoring in economics, statistics and corporate finance streams, has helped him to combine industry perspectives with analytics techniques, to deliver high impact projects. He has been able to apply practically over 80% of his academic learning.
• Typically in Philips consumer facing business like personal health, they combine insights from marketplaces like Amazon, data from their own e-commerce business, and google mobility trends, to understand customer segments, profiles & migration. This enables them to align supply chain and availability metrics, for enhanced growth opportunities.
• Before embarking on a project, it is necessary to understand expected outcome from Stakeholders. If the expected impact is limited and not strategic, the tolerance levels for accuracy is high and there are time constraints, one should decide to adopt a rapid analytics approach. The focus should be on using tools and techniques which can provide insights faster.
• In the health tech space, machines are largely IoT enabled and so there is a lot of data exhaust available. The key point in how to use it, should be determined. by the problem to be solved, and the impact of preventive maintenance.
• When it comes to collecting data, a good way to look at it` is to see the recency value of data, and the actionability within that time frame. Anything beyond that can be stored as summarized data. Another factor to consider when collecting data, is its at its relevance to already known problems.
• Business situations needing high levels of accuracy in insights or those governed by regulatory requirement, will need a higher level of explainability. In such cases, in addition to domain context, knowing the first principles of algorithms / tools being used, will be absolutely necessary.  However when the level of tolerance is reasonable, then good contextual knowledge , and ability to leverage open AI tools may be enough.
• Staying curious and ability to have a contrarian viewpoint and challenging status quo,  are skills that the next generation of data analytics professionals will need to mandatory have, given the rapidly evolving business and technology landscape.

I am sure you will find the conversation with Saswata very interesting. You can also watch / listen to the interview on our website, youtube, apple and spotify podcasts on the links below. Please do share your comments and subscribe / follow us on @maavrus.com on LinkedIn, facebook and twitter.

Today is Episode 2 of the Interview series on Expert-talks @MAAVRUS,  with Leaders in the Analytics, AI, and Transformation space.  For this episode, I am in conversation with Dr. Prashanth Southekal, Founder & Managing Principal of DBP Institute Canada, which is a data & analytics consulting, research, and education firm.  Prashanth has consulted for over 75 global companies including GE, P&G, SAP, and Apple to name a few. Prashanth is the author of 3 books – Data for Business Performance, Analytics Best Practices, and Data Quality. He is also an adjunct professor for data & analytics at IE business school, Madrid, Spain. He writes extensively for Forbes, FP&A Trends, and CFO University.

It was wonderful to speak with Prashanth and understand how he uses the education & training of corporate leaders, as a key step in the organization's analytics & Transformation Journey. Am listing below, some key learning from the interview.

• He sums up his overall approach to business as follows

1. It's all about building networks. You help some people and either they help you back or somebody pays it forward for you. So you always have to be part of the community and help people. 
2. The fundamental point is educating the business community about how data and analytics can be used for solving business problems, and it creates both the funnel and opportunity for him.

• Data analytics he explains is about Digitization & Data Management, Data Integration, Data Science, and Decision Making/execution. He believes that while Data Science is considered more glamorous, the opportunity to make serious business change is in data management, integration & practical execution.
• It is important to know clearly - ”What is the question one is wanting to answer” & “What if” before embarking on any data & analytics projects. His experience is that in 90% of cases, this is not defined upfront. So he says “No Questions” +”No Data” = “No Analytics”
• There is nothing called Perfect quality data, and businesses have to be comfortable with this truth, so “Perfection is the enemy of Progression”
• For businesses to make the leap from Performance Insights to Actionable Insights,  requires Business Resources, Leadership, and Openness to change. So he recommends that for data analytics to yield sustainable value, one must start with “the end in mind”. 
• For effective data analytics adoption, it is important to move companies gradually across the measurement continuum – start with initial descriptive analytics and move to prescriptive recommendation models. The biggest block to this is organizational inertia. Also jumping the queue does not give great results on account of poor measurement quality and decision rigor.
• Dark data is data that is not being used for any specific use, and it is costly to collect & maintain. So before deciding to collect data, it is good to check if they serve one or any of the purposes – help run operations, needed for compliance, and improve the quality of decision making.
• In the next few years, AI and Digital will increase the level of automation, but the role of good Data Analysts / Scientists will be to assess the quality of data, engage with businesses to help articulate the right questions given decision constraints, and enable businesses to act on recommendations. So the future of work as far as data analytics is concerned will mostly be influenced by soft skills, rather than hard technical skills.
• His new book “Data Quality” uses a framework called DARS, and uses 16 approaches to identify why businesses have bad / problem data, and 10 approaches to fix the data quality.