Ranjeet Bhosale Talks Analytics, Audience-Based Marketing and the Art of Storytelling

After completing undergrad at MIT, I did a couple years of software development back in India. Then, in 2007, I moved to the U.S. to get my MBA in finance. After getting my MBA, I joined Staples and did a few short stints in finance where I managed the entire business. Doing this, I learned a lot – including how to look at numbers, understand the health of the business and then tell that story. Where are your headwinds? What are your challenges? What are your areas of opportunity? How do you speak to the business’ leadership? Answering these questions and mastering the art of storytelling is a big skill set I developed there. 

With that technology and finance background, I moved into merchandising analysis to help merchants optimize their businesses online. From there, I moved into product management, optimizing the site experience and building new capabilities on Staples.com. I then had the chance to run and develop the entire pricing platform for Staples online. In that role, I led the global analytics organization where I managed all the different aspects of pricing, promotions, marketing analytics, merchandising analytics and online analytics. Every time I moved from one division to another, I had zero understanding of the role I was walking into – but what I brought with me was my past experience in technology, my understanding of how finance works and my experience with how merchants speak. All those things helped me a lot in terms of setting myself apart in my career.

I then had the opportunity to take what I learned at Staples and apply it on a completely different scale – making the shift from a $20 billion business at Staples to a $100 billion business at Home Depot. While the core capabilities and concepts are the same from retailer to retailer, the scale at which you have to apply them and the challenges you face are different – and that really tests you. After two years of leading analytics, I pivoted into a role managing the site experience for the decor part of Home Depot’s business. And, in the last two and a half years, I’ve been leading Home Depot’s audience-based marketing.

There are two key things. First, you need the core analytics concepts from your graduate program. That’s your foundation – that’s table stakes when you get into the company. But what will set you apart and allow your career to skyrocket is your curiosity and your knowledge of the business. This is what I’ve seen in my time at both Staples and Home Depot. The fresh graduates who really shine inside the organization are the ones who take time to understand the business. 
If you’re an analytics person who’s expected to provide insight, you can’t provide insight if you don’t understand the business. So do you know the business? Do you understand the trends? There are so many dashboards you can keep track of so that when somebody asks you a question, you have enough knowledge to provide proper insight. No matter how strong your technical skills are, you won’t be able to succeed if you don’t know the business. So have that curiosity – that’s what’s going to set you apart from your peers.

So let’s say that Annie and Rishika are both homeowners looking for home improvement products. Annie’s looking for ladders and Rishika is looking for faucets. How do they start the journey as consumers? If they’re like most consumers in need, they’ll either go to Google or go to homedepot.com to start searching. That act of consumer searching is a signal – Annie’s household is looking for ladders while Rishika’s household is looking for a faucet. And here’s how most retailers have traditionally done marketing in the last decade. They’ve said, “We don’t care what Annie and Rishika are looking for – we have an amazing deal on tools and appliances this week.” And so they start showing Annie and Rishika ads for tools and appliances. But Annie and Rishika don’t care about tools or appliances – they care about what their needs are.

What does this mean? It means that all the money they’ve invested into advertisements for Annie and Rishika was wasted. And to make it worse, marketers haven’t only displayed ads that aren’t relevant to a consumer’s needs – they’ve also shown them completely different messages across the channels. So they might see an ad for tools when they open Weather.com and an ad for appliances when they go to Facebook. What we want to do is break away from this traditional model and start showing households the content that’s relevant to them and make sure it’s consistent across all the channels. So whether Annie opens Weather.com, Home Depot’s home page, Facebook or Pinterest, the content that Home Depot is delivering to her must be what she cares about – ladders – and not what Home Depot wants to sell to her.

While I’m using Home Depot as an example here, I can apply this to all types of retailers – whether they’re in the home improvement space or not. So gone are the days when companies created one type of content for millions of customers – now they have to develop millions of types of content for individual customers. That’s the fundamental shift we’re seeing in marketing campaigns.

While this is easier said than done, there are four major pillars. First, we need to predict the customer’s needs. We need to have data science, AI and analytics at the individual household level in order to understand and categorize their needs based on the signals they’re providing.

Then, once I know that Annie’s household needs ladders, I need to start showing her content based on her needs. But because there are different types of ladders – 12-foot ladders for consumers, 7-foot step ladders and 25-30-feet ladders for professionals – how do I know which one to show her? One way to identify the right content to display at an individual level is to use AI and machine learning to look at the behavior customers have shown in the past. This is content-based optimization. 
Then, once we’ve figured out the right content to show, we need to figure out the right channels to use for individual households. Again, let’s use Annie and Rishika as an example. Let’s say Annie is a very avid Facebook user, but she doesn’t come to Home Depot often and she hasn’t subscribed for any Home Depot emails. Rishika, on the other hand, has a Home Depot within a mile of her house and is a huge fan of Home Depot because of its convenience. She’s not active on Facebook, but she uses Pinterest a lot. Even though Rishika has a Facebook account, we shouldn’t waste our money on a Facebook ad for her – that channel won’t be effective.

Finally, we have to think about measurement. Even if we get pillars one through three right, we need to consider the incrementality and business impact. Did my marketing content really drive the sale? Or would Rishika have bought the faucet from Home Depot anyway since the store is nearby? This is where A/B testing comes in.

I’m sure many of you are experts in SQL relational databases – so think about it like this. Every time a customer interacts with Home Depot, they give us a signal about what their need is. Through data science, we can stitch those signals together. So let’s say Annie visited homedepot.com looking for ladders on Monday. She searched for keywords like “step stool” and “12-foot ladder.” She looked at page one and clicked on a couple of assortments. She went to ratings and reviews. She checked the prices. Behind every single interaction is an intent – and being able to understand those intents allows you to make predictions. So look closely. What kind of SKUs is she looking at? Is she spending time looking at a particular brand? Everything she’s doing – and all of her past purchases – allows you to predict what Annie’s current needs are, how likely she is to make a purchase in the future and what her brand loyalty is. 

There are a few ways to do this – and let’s use appliances as an example. In our houses, we typically have a dishwasher, a refrigerator and small appliances like toaster ovens and microwaves. And most of the time, customers try to make sure that their appliances are from the same brand. So we can use a customer’s past purchases and browsing history as signals. If every appliance Annie bought from Home Depot in the last year was from that particular brand, there’s a very high probability that she’s loyal to the particular brand.

There are other signals as well. The amount of time Annie has spent reading ratings and reviews for particular SKUs, for example, is a good one. Using data science, you might give more weight to some customer interactions than others – and this can help you take a stab at whether Annie is brand loyal or not.

Now, let’s say you run a marketing campaign where you show a Samsung ad to a million people who you believe are highly loyal to Samsung. After running the campaign, 200,000 of those people buy something from Home Depot. You expected, according to your model, that all of them would buy Samsung products – but that’s not the case. Some people purchased LG products and others bought GE. Why did your prediction go wrong? What is it that you missed? You can use the new information as input to update your models and make your insights better and better. 

While we’ve been doing audience-based marketing for three years now, it’s still very cutting edge in retail – so several studies are being published about it. As the entire industry shifts toward the next level of personalization, analytics is still catching up to that shift and is trying to push the envelope by defining key metrics. 

But let’s take a step back. Almost 20 years ago, most retailers were doing marketing through generic TV ads and radio ads – and if they wanted to target consumers, the best they could do was use designated market areas (DMAs). At that point in time, region-specific marketing was really the only way to pursue personalization in marketing. Then came the digital era, where Google search and channels like Facebook, Pinterest and Yahoo became prominent. With this came a seismic shift in the field of marketing. Retailers started moving money from traditional TV and radio channels to these digital channels. In this new marketing landscape, retailers began to question how to measure their success. It took almost five years to come up with cost per mile (CPM), clickthrough rate (CTR) and other metrics. Knowing that marketing in this digital media space was more targeted and surgical, it took time to figure out the right measurement metrics for it. 

Now there’s another pivot happening. Marketers aren’t just talking about channels anymore – they’re talking about customers. Like before, they’re trying to figure out the best metrics, so several research papers and articles have been put out about this. Going back to Rishika as an example, marketers want to know whether their marketing dollars influenced Rishika to buy from Home Depot or whether convenience and brand loyalty led her to make the purchase. So some case studies are being created to determine these metrics and drive incrementality.

SEO is still very relevant – and it’s important for every retailer. Think about it like this. Each of us has probably searched on Google at least 10 times today. And as an individual, you know that you have more skepticism about the paid ads that show up than the organic content that Google ranks. And we’re seeing this at Home Depot. A significant portion of our web traffic comes through organic search – not the paid ads – which is a strong signal. Also, if customers need faucets, they may not start on Home Depot’s website. Many people start on Google. After seeing Home Depot ranked on page one, they click on it and that’s how they come to homedepot.com. That’s how we know that these customers are looking for faucets and that we need to serve them.

Here’s one of the philosophies that I really love about the founding members of Home Depot and even our current CEO – and it’s not intended to be cocky. They always say, “Don’t worry about what the competitors are doing. Focus on what the customers need right now and what they expect from Home Depot – and then cater to that. Everything else will automatically fall into place.” So we look at competitors in terms of how they are solving the customer problem – but we don’t get distracted by the new, shiny things they’re creating and we don’t try to mimic them. We focus on solving the customer experience. At the end of the day, that’s what the customer cares about – whether they walked into the store or visited the website. So whether you’re looking at it from a price and assortment standpoint, value proposition perspective or delivery perspective, it all boils down to customer experience. 

There are two ways to look at customers. One is to look at them in the rear-view mirror by considering how many transactions they’ve made in the past. The other involves predicting how many transactions they’ll make in the future – which can be a swing and a miss, especially in the home improvement space, which goes through cycles. One year, a customer will do a massive kitchen and they’ll have a huge spend – but they won’t do that same thing the next year. So in certain organizations and ecosystems, it’s very hard to predict the future.

However, this does not mean that these types of organizations shouldn’t consider customer lifetime value. What they can do is look at customers’ past behavior and use a technique called RFM (recency, frequency and monetary) analysis. It considers how recent their purchase was, how frequently they are buying things and how much they are buying. This framework allows them to puts people into buckets based on their past behavior and really understand their loyalty – which, in these kinds of scenarios, is better than trying to go by customer lifetime value.

However, this isn’t true for every single retailer. In the grocery space, for example, the average customer might spend $200 on groceries every single month – which can make it very easy to predict the future and know customer lifetime value.

If you just search “marketing analytics” on Google, you’ll see an amazing amount of job titles come up. Marketing analytics is a very generic description. And, as we saw before when we looked at the four pillars of marketing, there are so many specializations. You can do marketing analytics in media optimization, customer behavior prediction, content optimization, marketing effectiveness and incrementality and so on – but these are specializations that companies don’t expect you to have fresh out of university. They just need you to know the basics of marketing, have a curiosity about learning the business and an understanding of the different verticals. As part of a business, you’ll get to master one of these areas and have the opportunity to move across the different verticals and get a well-rounded experience. 

I’ll be very candid – prior to joining this role, I had zero marketing experience. Still, I was tasked by leaders with solving the marketing personalization problem at Home Depot because I had transferable skills. I brought with me analytics skills and online experience. So it doesn’t matter whether you’re starting in marketing or not. Get your foot inside the door of a company. Whether it’s online analytics, supply chain or finance, it doesn’t matter. Get into any of these verticals. Learn the art of storytelling by looking at the numbers and discovering ways to optimize the business. Those are the transferable skills you’ll need to move into the marketing space.

Start at the 50,000-foot level. Ask, “What is the company’s marketing budget? Where do they spend that money?” Then, go even further. Look at the numbers. Most companies have reports generated every single week and a dashboard you can look at. I can guarantee that it won’t take you more than 60 to 90 days to understand the nuances. And most importantly, ask why. Your college will give you the technical skill set – and that’s critical. But what you have to bring is curiosity – which is the hard part. Often, people get scared and worry that their question is stupid, which stops them from asking it. But 9 out of 10 times, it wasn’t stupid. It’s just that nobody else was asking it. So have the courage to ask the question and the willingness to learn. With that combination, you’ll be unstoppable in any organization.