Discover what's new at LatentView Analytics

As we move into a post-cookie world, here’s how data and retail analytics will change how brands and customers interact in 2023.

The beauty of being utterly connected to a brand, whether via in-person shopping experiences, Instagram advertising, or online customer service, is that every moment can be personalized and tailored to an individual customer.

For example, when you visit Sephora, you’re immediately able to have a truly individualized buying experience. You’re greeted with complimentary customer service offering free product testing and in-store tablets that grant access to your personal account with preferences and purchase history. Then at checkout, you can use points to access free products and even take home bonus items. Finally, after you check out, you’ll receive emails with instructions on how to use the products you bought.

The beauty of the marketing strategy—no pun intended—was in how it leverages data to build unique, omnichannel customer experiences. With a store that has hundreds upon thousands of options, Sephora offers its customers a targeted approach to shopping that taps into their preferences, history, and tastes.

Sponsored: Mainframe Data Is Critical for Cloud Analytics Success—But Getting to It Isn’t Easy [Read Now]


This type of comprehensive customer experience is the foundation of forward-looking brands. And without data, brands are flying blind! As we move into a post-cookie world, here’s how data will change how brands and customers interact in 2023.

1. Predictive Analytics Will Power Personalization

According to a McKinsey study, 78% of consumers said that personalized communication was more likely to make them purchase again in the future. But, as we move beyond third-party cookies, a new phase of personalization built on zero-party and contextual data is in full swing.

At the most simple layer of personalization, a brand can use transaction data and predictive analytics to proactively highlight relevant products when a customer is most likely to want them. For example, for household items that are repurchased several times throughout the year, like water filters or coffee pods, a brand with a strong sense of their customer’s purchasing behavior can push marketing communications at the exact time they are likely to make a purchasing decision.

This and other data from sources across the organization—when compiled—create a unique “customer yearbook” or history at-a-glance of a single customer’s engagement with the brand over time. Such data allows for regular, personalized 1:1 marketing and advertising.

For example, REI has consistently been on the cutting edge of using contextual data to highlight relevant products. In one campaign, they looked at weather and geolocation data to surface ads for ski or rain gear depending on whether it was snowing or raining.

Even without third-party cookies, there will always be insights from predictive analytics using privacy-compliant data. Personalization isn’t dead; it’s just being rethought.

See also: The Personalization Paradox and How to Solve It

2. Retail Media Networks: Automation, Collaboration, and Fragmentation?

Retail media networks (RMN) have the potential to fill the void left by cookies. According to the World Advertising Research Center, retail media ad spend is forecast to reach $121.9B globally in 2023, up 10.1% from last year, making it the fourth largest advertising medium.

RMNs are essentially advertising networks built within a retailer’s marketplace. Think of Amazon, where sponsored products now routinely sit above organic results. In fact, Amazon’s ads division was one of its fastest-growing business units in 2022, representing $31.16B in revenue.

RMNs are lauded because they give brands and agencies access to first-party data at the point of sale, where customers are actively engaged in shopping and are the most receptive to relevant ads.

This type of advertising is also easily automated. Many small Amazon sellers quickly calculate their bids for high-value keywords using simple Excel formulas, while larger ones are embracing AI-driven advertising.

Equally exciting are the possibilities with Amazon Marketing Cloud’s “Clean Rooms,” essentially privacy-compliant data collaboration zones where brands and agencies can anonymously compare their customer data with Amazon’s to find novel insights.

Sponsored: Mainframe Data Is Critical for Cloud Analytics Success—But Getting to It Isn’t Easy [Read Now]

As every retailer builds out their own ad network, the space will get more fragmented, and with that fragmentation comes data problems. Now more than ever, brands will need to invest in analytics to ensure that they can track and compare KPIs across different ad channels.

3. Experiential Shopping: Physical Insights Are Going Digital

As retail influencer Steve Dennis famously said in 2018 (and pretty consistently ever after), “Physical retail isn’t dead. Boring retail is.” We’ve seen that play out in the last 18 months as in-store shoppers have come roaring back.

But physical retail isn’t the same. It’s smarter, more digitally oriented. And it’s giving brands more data and customers more unique experiences.

With many in-store shoppers, grocery is leading the charge. In fact, according to Prosper Insight & Analytics’ “Prosper Wisdom Suite,” an insights-on-demand platform that includes zero-party data from a monthly representative survey, 77.5% of Americans have shopped in-person for groceries in the last 30 days, the highest of any category.

And so, it follows that grocers are the keenest on getting in-store data. In February, Wegmans announced it was testing out smart carts at two of its locations. This has the potential to reshape stores, quite literally. Using computer vision mounted on the carts, Wegmans can map customer routes to optimize their layouts. The carts also allow for contactless payment, improving customer experience and the ease with which they can checkout.

While grocery is certainly the leader, other industries are following suit, especially those with many in-store shoppers who like apparel and beauty. For instance, many brands have opted to build small pop-up installations within partnering stores or even as mobile pop-up trucks. They’re doing more than just gathering data—they’re using data to deliver unique, customer-centric experiences tailored to their target shopper.

Go Where the Data Takes You

During the pandemic, retail brands saw the importance of eCommerce. Then they learned the hard lessons of supply chain management as ships sat for weeks in ports across the world. As shoppers returned to physical stores in 2022, they saw the importance of omnichannel.

But through it all, brands were leveraging data like never before to make hard decisions. Demand is difficult to predict in this strange economy, but as retail media networks explode in popularity and traditional marketing techniques are being reinvented, the brands that use data to meet their customers wherever they will be the ones that win.

Anyone who has worked in sales, marketing or customer service over the past decade understands just how much these functions continue to overlap. As competition increases and companies expand their digital footprints even further into the globalized marketplace, individual departments are getting flooded with customer information from more touchpoints than ever before. This explosion of data has led to the emergence of a critical new team—revenue operations.

Revenue operations (RevOps) is the strategic integration of sales, marketing and customer service into a single powerhouse designed to drive revenue growth. By creating this centralized operational view, management can break down silos between departments, learn unique insights and gain a better understanding of customer behavior. This way, each department can manage its day-to-day processes while relying on RevOps teams to align their data, analyze interdepartmental trends and report on profitable new strategies.

Circa 2020. The COVID-19 pandemic was spreading rapidly across the globe, and the healthcare infrastructure in several countries was struggling to cope with the influx of patients awaiting the results of an RT-PCR test. It was during this period one frequently heard the usage of two words: false positive and false negative.

For the uninitiated, a “false positive” test happens when there is a positive result for an outcome that is supposed to be negative. A “false negative” test is the opposite of a false positive test. While these two terms are often used in the medical field, they also have implications in software testing, finance, and cyber security. This article will discuss the impact of false negatives and how to eliminate them using analytics.

The Trouble With False Negatives

False negatives can have severe consequences, such as missed diagnoses, missed opportunities, and increased risk. For banking and financial sector companies, if an incident or a transaction is flagged as “false negative,” it implies that no crime or fraudulent activities have occurred. Since no offense is committed, the investigators often overlook such transactions. And when such frauds are detected, they damage the banks’ credibility in the eyes of their customers and stakeholders.

Anti-money laundering (AML) compliance is a requirement for banking and financial institutions to ensure that the system is not used to launder “dirty” money. For example, imagine a scenario involving false negatives where a well-known bank’s security software fails to detect money-laundering activities during routine transactions. The result would be that regulators would clamp down on the bank, and the bank’s reputation and financial stability would be at risk. Worse, such incidents undermine the security and compliance measures the bank implemented in the system.

The prevalence of false negatives while screening for diseases had grave implications for the patients. During the early days of COVID-19, any individual who exhibited cough and fever symptoms was recommended to get tested for COVID-19. The test results returned negative in several instances, but the patient would still suffer from the disease. Such a scenario meant that the affected patients could not receive treatment and recover on time as their test results were negative. Not only were the patients carrying the disease, but because of the highly infective nature of the SARS-CoV-2 virus, they would spread the disease among their family and the immediate neighborhood.

False negatives in cyber security refer to instances of threats undetected by the security system, either due to ill-equipped security measures or the level of sophistication involved in the attack. As in the case of banking and financial companies, false negatives impact cyber security along similar lines. Massive data breaches, loss of intellectual property, ransomware infections, and so on are some of the consequences caused by false negatives.

The proliferation of mobile internet and broadband connectivity has boosted online shopping, and e-commerce marketplaces have emerged as the go-to destination for tech-savvy shoppers. Unfortunately, the convenience delivered by online shopping has given rise to the problem of false negatives or true positives. These occur when online merchants fail to detect or flag fraudulent transactions, and these bypass the merchants’ fraud detection systems. This is the opposite of false positives, where merchants decline transactions as they suspect fraud. In both instances, the merchants suffer due to loss in brand reputation, monetary losses, and lost customers.

The Helping Hand of Analytics

As mentioned in the previous section, false negatives can result in misidentifying and misdiagnosing ailments and diseases among patients, loss of reputation among banks and financial institutions, and exposure to security systems and cybercrime attacks. In such a scenario, analytics is critical in identifying and eliminating false negatives. Analytics is all about discovering and communicating meaningful patterns in data. It helps businesses make informed decisions on how the data can be used for deriving benefits, increasing sales, and reducing costs. Let us examine a few use cases for analytics in managing false negatives.

Time is of the essence for banks and financial institutions as analytics reduces the need for manual reviews, enhances the speed and accuracy of processes, and lowers the overall cost of operations. Analytics can handle a large volume of data accurately. Banks and financial institutions are inundated with a vast amount of data daily, which they need help processing. When these institutions rely on traditional data analysis methods, they often become time-consuming and error-prone. With its algorithms and data models, analytics can handle a significant portion of data and identify patterns or anomalies undetected by humans. In addition, Analytics is cost-effective and scalable in identifying false negatives.

The healthcare sector can benefit from analytics as it identifies abnormalities and changes, which could indicate the presence of a severe condition or a disease. With analytics, healthcare professionals can monitor vital signs among patients, which helps them identify the underlying causes and respond accordingly. Furthermore, analytics can eliminate the threat of false negatives in cyber security by examining network traffic, log files, user behavior, etc. It can also provide real-time insights into emerging threats and help quickly eliminate them. Additionally, analytical models can be constantly updated and enhanced based on feedback and data, which allows them to fine-tune their responses to changing circumstances.

Analytics can help e-commerce marketplace in identifying false negatives. Machine learning models with fraud detection tools and with sufficient rules inbuilt can easily classify and detect them. Another popular method in machine learning to deal with false negatives is implementing a decision tree algorithm. This enables merchants to deliver a warning to the customers, suspend them, or de-platform them permanently for indulging in fraudulent transactions. However, the biggest challenge in a decision tree is the changing and evolving fraud patterns. As a result, despite stringent thresholds, merchants may penalize at least 5% of the “good” buyers from their marketplaces.

Leading the Way

Businesses rely on vast amounts of data regularly. In such instances, the chances of false negatives are significantly high. Suppose left undetected or overlooked, false negatives impact business transactions and disease transmission, resulting in data breaches and loss of reputation. Analytics is a valuable tool for identifying and eliminating false negatives and is essential for effective decision-making.

Balakrishnan explains how data and analytics have the power to penetrate supply chain obscurities and allow insight into demand and supply variance

Please introduce yourself and your role…

My name is Sunder Balakrishnan, and I’m a Supply Chain Analytics Leader at LatentView Analytics, where I focus on helping global organisations build resilient supply chains by helping drive a “connected supply chain” vision using the power of analytics, AI, and supply chain consulting.

What are some of the most significant challenges that procurement professionals are currently facing amidst the global supply chain concerns, and how can they overcome them?

Procurement professionals live in a world of a constantly oscillating supply chain that swings between supply shortages and excesses. There is an ever-growing need for resilience required to deal with disruptions to global supply chains from the pandemic, natural disasters, geopolitical tensions, and other factors.

These events have led to increased costs, supply shortages, and regulatory compliance concerns. A traditionally cost-minimization focused function is now shifting its operating principles to think about the balance between cost and supply performance.

The dependency on China has called on enterprises to think about a China-plus-one / alternate sourcing strategy and nearshoring opportunities. These challenges point to procurement being at the cusp of a transformation from a cost reduction philosophy to service-level maximisation lever.

The first step in this transformation journey has to be a thorough supply-side visibility using a quantitative approach to supply availability risk assessment, and the associated cost analysis.

Proactive steps to diversify their supply chain, renegotiate contracts, find alternative suppliers, and implement cost-saving measures such as automation and process improvements would be subsequent phases of this transformation based on every organisation’s unique situation and needs.

The role of data+analytics+AI and technology paired with good old supply chain thinking will help organisations unlock their true procurement potential.

Can you share any real-life examples of how supply chain analytics have helped companies to manage demand volatility and improve their procurement processes?

A leading toy manufacturer had an interesting problem where new product demand was highly unpredictable with poor accuracy leading to high demand planning bias and excess finished goods inventory.

For a toy to be launched 6 months from now, the procurement order for parts had to go out today, and the process was entirely based on heuristics. The power of supply chain analytics with internal and external data factors, coupled with our consulting prowess, helped the organisation bring down the predicted demand variance, provide improved explainability to the factors affecting demand, and allowed the procurement team to have higher level of confidence in the BOM volumes required.

How has the pandemic affected the supply chain, and what strategies can procurement professionals adopt to mitigate its impact on their operations?

The pandemic has caused disruptions to global supply chains, leading to shortages of critical goods and services. These disruptions, as a consequence, led to increased costs for raw materials, transportation and logistics.

As the pandemic caused a wave of remote work, consumer behaviour and demand patterns changed significantly, making it a challenge for procurement professionals to truly assess and anticipate these new patterns.

However, procurement professionals, if they develop data-driven contingency plans, can achieve insights into factors that can impact the supply chain. Again, it goes back to that connected view approach that we have developed for our clients.

With the supply chain becoming increasingly disconnected, what steps can companies take to ensure transparency and visibility across their suppliers and partners?

Companies need to be able to view demand forecasting, supplier management and logistics management in a single-pane-of-glass view, understanding how they work together to achieve a full supply chain picture. By bringing together data insights from these pieces, which rely on data individually, companies can experience more connectivity between these individual hubs of data.

This requires data engineering and an infrastructure that supports this mission. It’s important for leaders to think about the future and not a matter of if, but when, the next major supply chain disruption will occur. So a capital expenditure today will save money and contribute to customer satisfaction in the future. The companies that took this step years ago to invest in data analytics are reaping these benefits today.

Every event or risk in the supply chain has a cascading impact on some other event downstream, but due to organisational silos, lack of visibility, sub-optimal processes, and technology silos, modern supply chains lack a good way to understand the impact of these risks on the eventual on-shelf availability. This is a disconnected supply chain.

Catena-X is a great example of auto OEMs in Europe coming together to build a supplier data hub for greater visibility and transparency to the supply ecosystem. With everything being intelligent today – from watches and glasses to swimsuits and toothbrushes, supply chains have converged and connected more than we realise.

How can procurement professionals leverage technology such as AI and machine learning to enhance their supply chain operations and mitigate risks?

Companies today can use AI algorithms to analyse data from various sources, including social media, news outlets and even weather reports to identify potential disruptions, including natural disasters, political instability and future pandemics.

Additionally, these tools can help procurement professionals identify the right suppliers based on their capabilities, performance and compliance with a variety of regulations.

AI and ML can also be leveraged to extract vital information from contracts, including terms and conditions, pricing and performance. These measures can help procurement professionals to develop contingency plans and mitigate the impact of these risks.

What actionable insights can you provide to procurement professionals to help them navigate the current challenges in the supply chain and ensure a seamless procurement process?

Diversify your supplier base so you have ample backups to work with if one of your suppliers is experiencing challenges.

Leverage technology to help streamline the procurement process with automation tools that can help reduce manual processes. This should include using predictive analytics to anticipate and prevent supply chain disruptions and have a connected view of your suppliers.

Foster collaboration with suppliers by building strong relationships with them to both understand their challenges and to communicate your needs.

Conduct regular risk assessments to identify potential risks and disruptions and then develop contingency plans to mitigate them.

Unsilo data and prioritise transparency both with suppliers and internal stakeholders.

Monitor market trends, including news, regulatory changes and other factors that can impact your procurement process.

Digital analytics consulting firm LatentView Analytics declared its fourth quarter earnings last week on May 9. Post earnings, Rajan Venkatesan, Chief Financial Officer, LatentView Analytics, elaborated on the growth drivers for the quarter and the trends shaping the industry, in an interview with Ayushman Baruah. Edited excerpts:

What were the key revenue growth drivers in Q4?

The revenue growth drivers in Q4 for LatentView Analytics were the broad-based growth across technology, financial services, and consumer packaged goods. The revenue from operations for the latest quarter was up by 20.1% to Rs 141.1 crore from Rs 117.4 crore in the same period last year, driven by the growth in these sectors. Going forward too, our focus will be on the value propositions that we offer to our existing and potential customers in these industries, addressing their pain points.

What were some of the financial highlights of LatentView Analytics for FY23 and Q4FY23? 

LatentView Analytics posted a consolidated net profit of Rs 155.4 crore in FY23, which is a 20% increase from the net profit of Rs 129.5 crore posted in FY22. The company also recorded its highest-ever full-year revenue of Rs 539 crore in FY23, up from Rs 408 crore in FY22. Profit before tax (PBT) for the full year was up by 44% to Rs 189 crore from Rs 131 crore in the previous year. However, for Q4FY23, the company’s net profit was down by 4% to Rs 34.2 crore against Rs 35.6 crore in the corresponding quarter of FY22. The operating revenue for Q4FY23 was Rs 141.1 crore, which is a 20.1% YoY growth, while the operating revenue for FY23 was Rs 538.8 crore, a YoY growth of 32.1%.

How are you impacted from the US banking crisis?

We have not experienced any direct impact on our business as a result of the crisis, such as loss of revenue or clients, as we do not have any direct exposure to banks in the US. Most of our work in the US is with asset management and fintech companies, who are largely unimpacted. However, there could be an indirect impact in the future, especially keeping in mind the likelihood of a recession. There may be a spill-over effect from its impact on other industries, who could feel the pressure of higher interest rates and access to credit.

How do you see the demand environment for the next few quarters?

Overall demand continues to be robust and the total pipeline continues to be healthy. However, sales cycles are getting elongated and there is some level of sluggishness in closing new opportunities. Our view is that this trend will continue for one or two quarters, post which demand should bounce back sharply.

Do you see any cuts in client budgets and delays or cancellations in projects?

We have been able to renew all existing contracts without any cutbacks for the next year, which is a positive sign. There is a consolidation of vendors happening in a few key accounts and we see this as a window of opportunity to gain wallet share in these accounts.

What are the emerging technologies you will focus on in FY24 and beyond?

Generative AI is on top of everyone’s agenda and we believe enterprises and service providers are both bracing for a significantly higher level of automation, on the back of more use cases that will get created with access to Generative AI.

What are your hiring plans for FY24?

We plan to add about 400 people through campus in FY24.

Given its distinct advantages, it’s a safe bet to assume that the connected vehicle will drive significant new cost optimization and revenue-enhancing opportunities in the automotive industry. Growth in connected vehicles and services is a critical differentiator for automotive OEMs of our times. It will be accompanied by large-scale connected vehicle data generation that can be leveraged to help save costs, optimize innovation, and to develop new products and services. All of this is clearly impacting the connected vehicle market like never before. A McKinsey report suggests that by 2030, about 95% of all new vehicles sold globally will be connected, up from around 50% in 2021. Overall, the connected vehicle market, estimated at $23.6 billion globally in 2021, will reach $56.3 billion by 2026.

Data analytics – the key to managing OEM costs

A key business value for OEMs from connected vehicles lies in its ability to save costs – and warranty costs in particular. Warranty costs alone consume 5% of the revenue that automakers earn, and the use of available connected vehicle data could cut these costs significantly. While many OEMs have been collecting data from their vehicles for years, it’s only recently that some of them realized the power of the data being collected and how to use it to actually reduce costs.

Sponsored: Mainframe Data Is Critical for Cloud Analytics Success—But Getting to It Isn’t Easy [Read Now]


Many OEMs gather remote vehicle data like speed, acceleration, distance covered, pause duration, temperature, etc., which provide a rich data set for further analysis. End-user driving styles can be identified using vehicle controller data and then compared with warranty claims data. These driving styles can be categorized into unique customer segments with different warranty risk profiles, characteristics, and costs.

By monitoring and analyzing connectivity data over a large section of vehicles, OEMs can detect technical faults and issues early on and address them proactively, which reduces the cost of repairs and warranty claims. In such instances, OEMs implement proactive diagnostics or prognostics, which identifies and predicts failures in vehicles. This is a win-win scenario as customers are alerted about potential failures or malfunctions before they occur, which enables OEMs to provide a satisfactory after-sales experience.

Groups that are at risk of high warranty costs can be identified through a process of dynamic profiling and continuous hypothesis testing. For example, if a group of 50 VIN numbers had some kind of pattern identified, which highlighted some unique issues like head gasket failures, a hypothesis testing tool can test aspects like valve failure when the vehicle was driven at a low speed to identify the root causes of head gasket failures.

Such hypothesis testing can identify clusters of vehicles, the underlying issues in each cluster, how the vehicle was used in each cluster, the potential risk of warranty costs for the cluster, the associated warranty costs, as well as the potential root causes of the underlying issues. Once the root causes and associated warranty costs are identified by driving style, predictive modeling is used to accurately predict future claims for each style based on past warranty claims data.

See also: How Connected Products Enable Predictive Maintenance

The overall business impact of the connected vehicle

The use of data analytics can lead to a significant reduction in total warranty costs. For instance, OEMs that have implemented such solutions reported a high reduction in such costs, which is a significant boost to the OEM’s bottom line. In addition, the analysis of data based on driving patterns leads to improved customer segmentation, which would not have been possible earlier.

Additional insights on how drivers use different vehicles like hybrid or electric vehicles also leads to additional insights into an OEM’s portfolio strategy. Such outcomes lead to improved engine development and testing, as the key insight from the data analysis is that short trips with long pauses cause higher engine wear and warranty claims – a very counter-intuitive argument. This enables OEMs to make better decisions about component selection, for example, by verifying whether there is the use of appropriate cooling systems in different climate zones.

However, a large number of OEMs are still new to the process of developing their digital strategies and infrastructure to enable connected vehicle applications and services. Tier 1 suppliers to OEMs are further lagging in most cases. Yet, offering connected vehicle applications and services can lead to significant reductions in operating costs (e.g., warranty) and topline growth (e.g., new subscription models).

To address this challenge, many OEMs are rapidly building their data science and engineering teams to be able to gather enterprise-wide data and use that data effectively to provide these services and applications. However, there is a war for talent – and the shortage of qualified data engineers and scientists is one of them. New pure-play data science and engineering service providers can help fill this talent gap.

Given its distinct advantages, it’s a safe bet to assume that the connected vehicle will drive significant new cost optimization and revenue-enhancing opportunities in the automotive industry. Of course, there will be challenges, especially in terms of how tech advancements will fit into customer requirements as well as the ability of OEMs and suppliers to effectively execute their digital strategies. But nothing will stop the connected vehicle from enabling a hugely successful growth story for the industry in the coming years. It is truly an exciting time in the automotive industry!