Beneath the iceberg of Customer Retention lies Customer Understanding | #223 Veda Konduru

Please support the show if you liked today's episode:

• If you love the podcast, please get someone else to listen, too!
• If you enjoyed this episode of the Ecommerce Coffee Break podcast, please head over to Apple Podcasts, leave a rating, write a review, and subscribe.
• Share the podcast with your family, friends, and co-workers.
• Tag the podcast on Instagram @clauslauter and let me know what you like about it.
• If you like the content and would like to support the podcast, you can buy me a coffee here.
• Become a guest on the show or sponsor an episode.

Claus Lauter: Hello, and welcome to another episode of the e commerce coffee break podcast. Today, we want to dive into artificial intelligence, AI. How can it help you as a merchant with your day to day business? Now, as a merchant, you have pretty much the same questions coming up all the time. So you want to find out how to optimize traffic, customer acquisition, increased conversion rates.

Maximize customer lifetime value, retention, increase average order value, bring back lost customers. All of these questions go through your mind. If you're running a business and there is a solution or there are solutions out there that will help you with that. And as an expert on this topic, I have Vida Konduro with me.

She's the CEO and founder of VectorScience. com. Prior to her entrepreneurial journey, she has been over a decade plus of experience in the corporate US as a. Technology architect as a programmer, as a business analyst, and with her academic strengths and strengths in stats, in math, data science, and programming, coupled with a professional experience in the additional functional areas of demand planning, forecasting, inventory management, and so on, she has started her venture into the AI space.

There she excels with tech product innovation, vision and strategy, product architecture and solution blueprints, and data science mythology. So she has a vast background when it comes to doing all these calculations on find out how to make your business better. And we want to dive right into it. How are 

Veda Konduru: you today, Veda?

Wonderful. I'm doing well. Thank you so much, Klaus. Thanks for having me on your show today. Good to have you on the show. Thanks for the kind introduction. Thank you. 

Claus Lauter: Let's dive right into it. So talking about artificial intelligence, helping you to make your business better and to basically have a tool that helps you answering all these questions that you have when it comes to marketing, to acquisition, to retention, give a bit of an overview, how you see the best entry to deal with all these different topics.

Veda Konduru: There is a very powerful connection between the relationship between acquisition and retention in today's topic. I would like to expand more tuned on the customer retention and why customer understanding is important. But then quickly looking at the equation of the acquisition and retention.

Many research studies have shown that Increasing even 5% of, customer retention can increase your profitability anywhere between 25 to 125%. It's startling. Bain Company has done this research and the corollary to that is it takes 5X, 5 fold. Five times more costs when you have to acquire a new customer.

So all the more reasons why it is important and it is the most sought after aspect in any customer centric e commerce business to go after customer retention. So I would like to talk, it's more of an iceberg when you think about it. We look at customer retention as are we getting sales the profitability growth, the leaders at the top board leaders board members and leaders.

The predominantly interested in the overarching scorecards, but then the makeup of that is a huge story. There are a lot of calls in the machine that come together for customer retention specifically. And every e commerce can benefit from this. They don't have to lag behind. They don't have to do the guesswork and targeting.

It doesn't have to be rocket science or it's not a far fetched story. But my story today or the perspective that I want to bring to your audience today is about why customer understanding is at the bedrock of bringing the best transformative customer retention results. as I use this analogy, beneath the iceberg of customer retention lies customer understanding.

So the tip of the iceberg, as we talked about, are the revenue profitability, if we conceive that. So let's break down that iceberg or the keywords here. We talked about why customer retention is important and how it plays a role in the equation of acquisition to retention. So what is customer understanding?

So in a loosely defined term, I could say it's the ability to understand the entire customer base by various criteria based segments. Let's break it down. Ability to understand the entire customer base. No matter how big it is, a million customers, a 2 million or 10 million doesn't matter.

It's not the scale of how many customers that makes it unreal day. It's just the mindset of, not applying the right tools here. So how do we understand all of them? And what do you mean by understanding? There is no one set of definition or a package that you get it and say, I got all the customer understanding.

It's ever evolving. how do we define it? So I use the word segments, right? Criteria based segments, I will go into expanding that in a short while, but segments right let's talk about segments. It's a very common definition, right? So a set of, customers collected in one place who bear the same common characteristics by the parameters of your choice, like you say, whoever is lives in this region, it becomes a segment.

 That's a very simplistic view of what segment is. But then here we are talking about a more, intrusive and a more deeper one, which is Also the collection of facts and qualitative attributes at every individual level. So here I'm bringing the concept that while segment is what we care about at a rolled up level, we are going slowly deeper into the iceberg now.

So what makes up a segment? It's a collection of customers now of the characteristics, but we need to build the characteristics of at each every individual level. We need to collect them. But these are not the shallow or the superficial ones like. The recency, frequency, monetary value, or the total spend, or what product they purchased, when did they purchase last.

These are very, very superficial and they cut very narrowly to your understanding. So the understanding that we're talking about is quite significant and powerful and enlightening in a way, because as we go down we can. Understand customers about their interests, motivations, behaviors. We keep hearing these keywords all over the place.

This is not the first time, but what do we actually mean by that? whenever there is a communication, digital communication happens. It's a signal happens. It's a virtual communication between the customer and your brand. So the interest could be. Product, what products interest them, what is the most purchased product category, right?

What is the least purchased product one? You can do something about it if you know, or if you ask the question, correctly to what matters to your business and the policies, right? And then you match it to asking that question correctly in segmenting them and understanding, Oh, X number of customers purchasing this product more.

Okay. So I need to have more inventory or more supply. So it leads, it's relegates itself into a very good virtual cycle of asking right questions and it goes into the continuum of the, entire business operations, not limited to this. Right. 

Claus Lauter: Yeah. Let me ask you, there, where, when, where do you start your journey with your questions?

Obviously, as I said, you come from a high level from the sort of superficial questions Yeah. You. Make your way down. What do you see are the most common questions that merchants should not ask? 

Veda Konduru: Should not ask is about what everyone else is doing.

There is no canned version of, , interests or motivations. What should interest them? If I run a promotion, for example, is it leading to a sale? When you make some promotional offers it does some sales, but then it is about understanding a customer segment.

 Your good question would be, Tell me customers who are making a purchase only when I make some sort of a promotion , and then tell me who are not making a sale, even with the promotion and tell me who are those who do not care about that. And then they purchase at a regular price.

Now, these are the questions you could ask, but then what they should not ask is, what is not aligned to their customer. Marketing efforts, because many times I see that marketing design or the campaign of it has its own rhythm and it has its own path while the segment has its own, it's an afterthought.

It just patched up at the very end. And of course the results won't show up because they have, there has got to be a continuum of, okay, I need to build this marketing campaign. I have a new product introduction. I need to understand who are my early adopters. You cannot ask a question of who purchased what not to ask who purchased in the last last three months, because I want to run a campaign for my new product introduction.

That's a wrong question, 

Claus Lauter: It makes total sense. And I think that's one of the very common segments that you see in a lot of tools. And sometimes they are also like pre filled in some tools is like the last 90 days. And I was wondering is like, yeah, what does that actually tell me?

I mean, yes, they bought, but I don't know the reason why they bought or why they were specific or what was the trigger for that. So I totally understand that makes total sense. 

Veda Konduru: I will expand on what you exactly said. So when you do a rule based like that. The danger to that is that the customers are not one monolithic group, right?

Customers as a broad buyer persona, they flock to a certain brand as they bear certain common characteristics and they become your customers. But very soon they are unique in their buying journey of what product they are, they have affinity to, right? What experience made them stick longer than the others?

 Or it is tied to their life journey, right? Life has moved on. They moved to a different house and now the products some of the products that they needed earlier, they don't, they no longer need, or they need more special occasions, birthdays. That alignment is not telling the 90 day rule of what they purchase, right?

It's because they have their own story and their journey in life and their buying journey is reflected in their purchases when they are coming and honing and cooking and opening and making a sale. We are not robots. So we have to tune dynamically to their story. That's what the understanding is about.

Rather than. I have this campaign to run. Let's get some combination of customers, just let's find some customers IDs in the system and just, send it out and just feel good about ourselves that we did some campaign because it's not going to get results. 

Claus Lauter: Yeah. Very good point. I think a lot of marketers unfortunately work in that field is like, I need to produce some results there.

And I think highlighting really the customer understanding is important. Now AI comes in when it comes to the pure volume of data that is coming in and then still understand what's happening. So how do you help with 

Veda Konduru: that? The beauty of technology, how it comes is it's an ally. It's a complimentary thing that meets the mind of the leaders.

So someone has conceived an idea that he, for every mark, a marketing campaign, I want to do a personalized campaign that goes to the right targeted ones. So with the technology that we have at vector signed or the technology at large, the power is in in creating that individual story.

Ask n number of attributes. How many ever you want to create? Like, are they likely to repurchase in the next 30 days? What product are they going to what cross product are they going to purchase? What is that longevity? How long are they going to? Because is there a need likely to churn score? For example, if you know you can do something about running a promotion for them, because historically, if they had a favorite product, then you can give a clearance to that and create a segment of it.

 It's a push and pull, and it's a dance between understanding the customers and then knowing your strengths of your product and inventory, your own, objectives of how to increase the revenue. They all come together with the power of AI. We bring the predictive insights.

So let's say there are 50 different things I can produce with the AI technology. I do a mutually exclusive signal saying that this particular customer A, has this engagement index. It's, he's been declining in the engagement, engagement in terms of clicking my campaign but he made like 500 in his lifetime, 10 purchases, but now, his motivations, whenever there is a clearance, he can, he'll buy.

And this is the product XYZ that he's interested in, but very interestingly, he also purchased product, N lately. So you suddenly, when you have a story of these 50 attributes, for instance, right, for simplicity's sake. Imagine yourself in your mind, a sort of a grid with all the millions of customers and all the attributes mapped out with their values.

You are enormously powerful and equipped to build micro segmentation of your choice. So the marketing executive come and ask, okay, here are the things that I want to understand. What are the customers who did XYZ? Okay. It's not a thing that you go and build a report and cook right from the scratch.

You have everything there. They go powerfully and then they build you run a query against all that meet the criteria. And then here, your micro segment, the power of segment of one also comes from peer class, right? So what we do is. Understand every customer by their strokes of these individual insights that includes predictive, derivative, computed the superficial ones, the deeper ones that tell a story.

Now you have the power. It's like a potential. Now it may, it becomes a dynamic energy becomes a dynamic power when the marketing executive says it is a marketing campaign. I want to run. And give me the segments X, Y, Z, or we can match to personalized campaign saying, Hey, you're running an NPI campaign for this product.

And these are the related products that really performed well when you release the product within the 3 months. This is the customers. Now, there is a lot of nuance class across what we do is level one, understand the customers with that deeper story that I told, bring the customer segments of your choice as you want dynamic customer segmentation for personalized marketing.

Number three, you can tell a story for every single persona, like for every single campaign, right? So you can make it more powerful. Ultimately, it has to lead to strong customer relationships to instill a sense of loyalty to evoke the emotion. For example. Meyers, when they send this handpick for you, despite me being the technology person, I know what happens, it feels magical when you can connect to what is offered in the coupon book, right?

Because it, it tells me that. The brand cares about it. The retailer cares about me, They know what I purchased. They know what I don't purchase. It tells a story. It's like a friend, millions of customers. With the technology it's possible.

And that's exactly what we do, what Amazon does. Customers who have purchased X also purchased Y. That's a product market basket analysis. So we do that as well. 

Claus Lauter: The word understanding for our listeners is coming from two sides.

You as emergent understand your customer better, but your customer gets the feeling that you understand them. And I think that's the most important thing is that they feel be cared about, understood that brings a very close or builds up a much closer connection between a customer, a client and emergent or.

A someone who sells online now, tell me at vector side, obviously with the solution you have is how does the implementation work for instance, if you're a Shopify merchants, how do I get the data in and how does it look in my day to day work as a marketer? 

Veda Konduru: For a vector science use case, we do cross a lab scale upsell scores, predictive scores, the segmentation by choice.

Right, the power of segmentation. So the attributes that we can publish, these are all the various details available at every customer. So they can pick and choose and create dynamic segments, but we also have the canned strategically profiled insights that we deliver by dashboard. So that's on the delivery side.

So coming to the, what data is needed, there is a unique thing that we do, which is combining the purchase history from the ERP or the Shopify store with the marketing campaigns that they run using Zendesk or any other Cleveo MailChimp or any other software that they have, Zendesk is for customer service.

 What we do is also as part of. That so one, we take that data, we ingest that, historical sales data via APIs historical orders and the customer address book or the customers who are your customers. We don't need the PII at all. We just need the ideas about the customers is fine.

And maybe the regions, if there is an element of region as a criteria and then the campaigns. So what we do specifically in our software. Is combine them, overlay them as a chronology to tell a story and bring you insights about you send four campaigns and the fourth one turned into a sale and then you did not send any, but the sale happened in this case, but in the other case you send 10, but they did not return any sales.

 There is a chronology and then we build on the technology side. We talk everything in terms of scoring because it's continuous as opposed to discrete. But then the business, can create labels by bucketing them into, if these are the scores, right, I can segment these scores and then call them rising stars or losing game or I don't know.

No, so it could be, those are the labels at the very top. So how it works for our software is it's not a hundred percent SAS, at the moment it is for enterprise e commerce merchants. So anyone with, with a decent history. So when I say decent history, it's about the volume. So it could be one year, but then they're just viral and they have, 25, 000 or more orders in the database.

So that should be a minimal volume of data to even do the ML on it and bring reasonable, insights. So there is that on the cap of it, on the data volume purchase history, campaign history are minimal, and you can add social media or the customer service as an additional databases from the disparate data to connect the story depending on what they want.

 But the standard one is about purchase history, campaign history. We'll give you cross sell upsell scores, predictive scores who are likely to repurchase what in the next 30 days, and then we have an API back to their ERP, which is Shopify. So they can collect the customer tags is what they call Shopify.

 They can collect all the segments of their choice and the tags, and then schedule the marketing campaign saying that this segment, this is the message and it goes out. And then you and then KPIs and then the KPIs, right? So for a long time, I was in this the solution business, the capability, but then I believe that the biggest fulfillment to me is when it makes the traction.

So there is a sense of KPIs. So if there are standard KPIs, there is a sense of what KPIs are in the direct line of impact, which is average revenue per customer, the direct sales, of course, right? And then second time to purchase. How quickly they're making a purchase. , if the customer or the merchant doesn't already have these KPIs, we can help them benchmark them, give them the methodology, how they are built, use that same token pre and post to show how well they are.

Performing I have seen some customers who are not, intellectually honest with themselves in the benchmarking because they want to see something really good to sell. But that's the very minority. Let's not talk about the minority here, but people who are really serious about growing the business and want to know the truth what's going on.

I want to change. This is one step closer or several, steps closer. It's a good stride. In the right direction to get us transformative growth. 

Claus Lauter: Tell me about the onboarding process. Is there a kind of homework that the merchant, needs to do before they can start onboarding?

And then how long does it take usually? 

Veda Konduru: Yeah, so it takes about, it used to be a longer cycle, outside for the enterprise with a lot of customizations. We have shrunk that timeline. Now within a week to week and a half, we can get them up and running. It's a pretty significantly short time compared to why.

It's relative one can debate if it's all 100% you turn on yourself versus that's a canned version. But in this one and a half, what happens is what is important because it look, it is a custom tailored. We can ask some interview and ask some questions, good questions to the marketing executors what matters to you?

What KPIs are, bothering you? what kind of questions would you ask? I wish I asked or I wish I had this insight I would have done. What is that power that you want for you, right? And we can quickly bring that all of that to fruition by building. It's not building our models. We have pretty pipeline, the AI pipeline.

But we have the ability to configure. And test and back test you to see how well they'll perform on the field for the data landscape because it has to do with the what they're selling if you're saying that the customer ideal customer profile that you're talking about today is commodity items repurchase required.

Consumables, for example, right? Repeat purchases. There is a sense of repeat purchases. Even within that, the price point can throw this, thing off where a particular niche one sells at 120 is the minimum versus a merchant that has anywhere accessories from 5 ranging to 200. there is a more customer base, a more frequency, a different way of understanding the patterns and the landscape.

So in this week, we do that tuned custom tailored API, driven both ways in data ingestion and on the outbound giving back to them. And we also are invested in working with KPIs, how well they're doing before and after from time to time. Okay, sounds good. Plus, to clarify, we were on Shopify right now.

We are not a app or anything to turn on. We are an enterprise one. But there is always an opportunity to reach out and connect with us for the enterprise ones. Traditionally, connect back to Shopify with the authentication, the right authentication from the merchant, from the client, we can always send our APIs back to them.

I don't know how close to that system now is, but that should really enable them to see the power of what is out there to improve the marketing. 

Claus Lauter: Excellent. Give me an indication about your pricing structure. How does that work? So 

Veda Konduru: there is especially in e commerce, it is priced by for a customer count because,

because we believe in that philosophy that customers are not ideas and you're telling a story when they come and then go, you, it matters and then it is by the stroke of how many customers we analyze, and it is a very small token. I think it is about 10 cents, depending on on our website.

So for every thousand customers. It doesn't linearly scale. If you have millions of customers, it doesn't linearly scale and goes to, it doesn't go to 10, 000, of dollars. It has to be palatable. We'll work with the higher volumes. We'll work with them. If there are adjacent insights that they are interested, then what we are offering, we are open to releasing them.

So there is a lot of comfort. It's as if they have their analytics and data scientist team in their own company. That's the comfort that we give because ultimately I believe in the movement of the needle than people making money and trade. That's not where my fulfillment comes from.

So the whole point is to. So we are flexible, but when it comes to pricing, it's on the website, but we can always look at the higher volumes. What do we do with 

Claus Lauter: that? Okay, cool. I will put the links to your site on the show notes and you will be just one click away before we come to the end of today's coffee break.

Is there one final thought that you want to leave our listeners 

Veda Konduru: with? My utmost interest to every customer centric e commerce customer is my heart is there and you next time. How they should think the long term, the long game as opposed to the quick fixes and getting the marketing or ROI and in the grind.

So that has to come with the forward thinking of I need to seek the truth from the data. Let the data teach me what it has been doing, what has been going as opposed to going with. I know everything. I've been in this business for X number of years. I can tell, who are coming, who are going, we don't need to do analytics.

The kind of mindset is obsolete and the sooner each leader comes and embraces technology for the right reasons, the more humans will be empowered and the human intelligence is at its finest and best only when technology is put to work at for what it does the best. So don't need to fear tech.

Be the forward thinkers and the critical questioning. Everyone needs to question everyone else in the organization. Why are we doing this? What are we getting? What do we not do better? Can we ask better questions? How can they make AI practitioner or a vendor like ourselves accountable?

can we be honest in taking our accountability and giving the clean data right data? Sharing the right information, right? And then make them accountable for what they are promising. So someone has to be seriously invested in the long term and the strategic and the deeper area of life.

Where is it making the traction? Because every day ghost come day comes and goes, and superficially if you do it and shallow life comes and goes. But the traction happens when we touch it a little bit deeper. Cutting deeper. Yeah. 

Claus Lauter: Very true words. Just following your gut feeling is not always the best advice.

Veda Konduru: There is a connection. Plus there is a connection. Take that. Don't discount your gut because it's powerful. But don't run with it, right? Don't execute on it directly. There is a methodical approach. There is a method to question that it's treat that as a hypothesis, right?

If I have a gut statement, then take that as a hypothesis. Don't get too attached and in love with it so much that you cannot detach, right? You have to be in the mindset of that curious mind that says, Oh, I had this hypothesis. Oh, it failed. Oh my gosh. You should be able to laugh at it. Right.

And then say that, Oh, this is what I've been thinking for years, though. This is not the case. Okay. Now I understand why things have changed. I appreciate what has changed, why has changed now you take on, and then you become more powerful with the different strategy, 

it may lead to a different policy. It may lead to a different product making using your own strengths. Who knows, if you're not asking the right questions and a gut has to be a hypothesis. And an experimentation, it has to follow the methodical approach using the technology asking right questions and it has to come to a fruition for the results are the best.

Claus Lauter: Yeah. Let the data guide you and you will have the right outcome. Thanks so much for your time today. It was great to look into what you can do with AI and how it helps you in your business to find the right points to make the right decisions. Thanks so much 

Veda Konduru: for your time. Thanks for having me, Klaus. It was a pleasure speaking with you today.

Thank you so much. Bye. Bye.