Transcript
Welcome everyone and thank you for joining us today. I'm Ayan Shahar, director of strategic partnerships at USD Spark and I'm happy to be your moderator for today's session on domestifying AI AI powered pricing management. Pricing is no longer just a financial decision. It is a strategic function that drives business growth, customer engagement and long-term profitability. As the market evolves, so do the challenges organization face in optimizing pricing strategies. Today, we're here to break down the complexities of pricing management and explore how businesses can adapt to stay ahead. To guide us through this journey and pricing is indeed a journey, I'm happy to introduce our expert panel, John Taylor, retail leader, UK at USD.
Taking over 25 years of experience in retail operation, distribution, and contact centers into global technology organizations, John now revels in bringing it to life cutting edge technological solutions that drives real benefits across the retail value chain. Hi John, thank you for joining. Um, also we have Dr. Fabian Urik, chief product officer at Quick Lizard. So before joining Quick Lizard, Fabian gained extensive experience in retail pricing as partner and global expert for pricing at the Boston Consulting Group and as head of pricing and online marketing at Zuplus, Europe's largest online shop and pet supplies. He holds a doctor's degree in behavioral pricing from TU Munich and is a resident guest lecturer for pricing at ETH Zurich. Hi Fabian, thank you for joining. We also have here uh Shrihasa Udupi a lead enterprise architecture from the UK um from US.
Shrihasa ensures technology investments align with BIRA strategy delivering scalable and future ready retail ecosystems bringing deep technical perspective. Shrehasel will walk us through how organization can integrate pricing intelligence into their enterprise architecture for scalability and efficiency. Hi all, thank you for joining us today. Throughout this discussion, we encourage you to engage, ask questions, and share your thoughts using the chat. And as a special takeaway, at the end of the webinar, we'll provide you with access to a pricing maturity questionnaire. This tool will generate within minutes a personalized report to help you assess your current pricing status and identify areas for growth and improvement. Okay, so let's dive in and John, we're going to start with you. From your perspective, what are some of the key challenges that are facing the retail industry today?
Thank you, May. So, that's a broad question and I'm going to try and pick some things out just to dive straight in. So, let's think completely away from pricing for a second. Supply chain costs. Everything about supply chain is seemingly more of a challenge. So whether that's the costs themselves, visibility, resiliency, maintenance, but in particular the challenge of making sure you can get a return on things like investment made in automation in warehouses and distribution centers. That's a really important part of some of the challenges that that retailers are experiencing today. Outside of that, optimization of space utilization across retail estates. We're seeing, you know, bigger stores that in in changing the way that people shop today, they need to be fully optimized.
And, you know, the introduction of concessions for synetic synerggetic partnerships and the introduction to physical digital media will really help utilize that space better. We're also seeing retailers expanding or looking to expand into international marketplaces and and you know, trying to work out how best to enter those markets and and make it a success. Omni channel that hasn't got that hasn't gone away. So omni channel integration and hyperpersonalization is um something that every single retailer is really looking to try and achieve. Um it it makes them more efficient. It allows them to get more insight into what's happening in in their environment and with their shopping habits. Um and then the last big rock if you like is the broad issue of costs challenge and profitability.
So we know that retailers are under profitability pressure not not least due to things like national insurance here in the UK the employer national insurance contributions and you know there's a recent survey with over twothirds of CFOs suggesting the prices might have to increase to compensate. So I think you know there there some of the big rocks just to answer your question Mayan but but actually in order to address some of these profitability challenges pricing we know remains one of the top three conversations in the boardroom today. So Jonah I maybe I need to start or continue with a provocative question based on everything that you just said and there are so many challenges. So what why should it be a priority for retailers to address this issue of pricing and I would ask me the same question. So so pricing can touch and influence so many strategic measures.
So revenue, profit, productivity, market share, customer experience and most importantly trust. Um, and depending on those drivers, we've seen retailers experiencing revenue lifts of up to eight%, profit increases of up to 5% and productivity improvements of 3x on the back of implementing robust AI powered pricing management solutions. So they're quite powerful metrics and I think hopefully that answers the so what and Fabian I want to take this next question um and ask you how would you describe the current maturity of the market when it comes to pricing and how organization are approaching and managing their pricing strategies today what do you see in the market so let's have a look at the pricing maturity framework from Gartner which spans across five stages from purely manual pricing to fully automated AI powered price optimization.
Most of the retailers we see today are on level three which means they take few selected factors cost obviously and most of them also considering competition but run the calculations in a more manual mode leveraging Excel or other calculation tools and the D2C brands or brands selling D2C actually we see even one stage behind so most of them that we usually um see in the market are on level two so even more manual and even less automated um for the time being and Fabian given um the market challenges that John described with supply chain cost visibility um resilience and so on maintenance how can pricing help? So basically all these factors influence uh the three core elements of pricing being cost the market and competition and customer and their value perception.
So looking at those um pricing of course needs to make sure that the cost are covered from the product cost but also supply chain and all the delivery um cost associated to bringing the products to the customer. Um second when we look at the value dimension it's important that price is seen relative to the value being generated by the products and therefore also here we see all these factors influencing and being very volatile these days and third looking at the competition and the market in general and also here we don't act in isolation so price always needs to be relative to the competitive alternative offers.
Now thinking about manual pricing you can imagine it's very hard to manage all these factors at the same time and usually you end you see companies ending up focusing on one or two of them and like waves maybe passing on the cost in one wave and then making sure they are not pricing themselves out of the market and this is actually where AI really comes in and makes a difference. So leveraging AI you can balance and manage all these factors simultaneously in at the same time. If you go through them one by one, for example, article segmentation is is a module that can help identify the different roles items play for the shoers psychology and therefore impact a lot how prices should be set. Or looking at competitor sensitivity and vicious pricing cycles that can be detected leveraging AI models.
This of course helps a lot to stop a vicious pricing circle um downwards and make sure that margins stay healthy. Understanding AI or leveraging AI to understand price elasticity and cross elasticity is key to make sure when anticipating or let's say when setting prices to anticipate the the adverse effects on other SKUs um and also the expected reaction of customer demand. And the next one forecasting and simulation basically this is one that brings them all uh together and makes sure that whatever price decisions are being taken they are considering the on let's say the future um impact of that and actually also you can optimize uh based on the expected future reaction of demand and maybe only one last one on the cost um which which I left so far this is often a very simple and isolated use case using AI to optimize for inventory um optimizing marked on clearance sales.
And Fabian, can you double click on the article segmentation? Can you elaborate on that? I think that's a that's a good idea because also this is the one that probably best describes the the theme of today's call demystifying AI based pricing because this one really helps to break down the complexity of of AI based pricing by modularizing it in the application. So articular segmentation basically means identifying what are the KVIS which are the SKUs the products that are the reason to go shopping that drive traffic into the store and for those SKUs it's really important to have the best price at every moment. The other end of the scale which we call the profit generators these are usually items at the end of the shopping journey uh sometimes impulse buys. Imagine yourself in a supermarket. You did your grocery shopping and then you find some batteries and flowers at the checkout area.
You put them on the belt without even thinking about the price. So here we can optimize leveraging I for um for margin and profits in the end. And sales drivers is a mix of both. So customers are still price sensitive but not as much as for the KVIS and therefore prices don't need to be on the daily base matching competition but they can stay on market average but give enough room and allow for temporary promotions. Now making this a bit more visible and tangible in terms of the pricing strategies that the AI models in the background would will optimize for. Um these are for KVIS competitive matching for the sales drivers um finding the revenue maximizing price point within certain competitive guard rails and for profit generators.
This is uh eventually finding the best opt profit optimizing price point that gives you the highest returns and this actually is done by I automatically in the background. So we run a model using more than 20 different metrics that identify the different roles of the SKUs and assign them to those three segments based on the price elasticity based on the basket position as mentioned before with the flowers and the batteries at the checkout based on the rep purchase rate purchase frequencies. So multiple factors that feed the model to identify the respective role and assign the respective pricing strategy. Fabian, can you bring it to life to us? Can you um explain what are what is the value in this approach?
So, specifically on the article segmentation and competitor sensitivity, which were two modules we implemented with Sephora, a UK based health and beauty retailer, and we really achieved astonishing results. Um on the one hand we identified that some of the competitors they've been always matching um for um are not as relevant are not in the relevant set of shoppers of their shoppers at least and this came out of the um competitor sensitivity model finding the cross elasticity between price changes of competition and the demand on their own um shopper base. And second uh the article segmentation helped a lot by identifying some of their KBIs which they thought they are super important to be price aggressive but in fact they were only safe drivers.
And therefore we were able to uh take out the aggressiveness and earn more money on these SKs without jeopardizing um the the turnover and and the units sold. And this was done in only three months. um and also increased the the key business metrics like profit went up by 25% in the test group versus a control setup. Um we were able to um reduce manual workload by more than 20 hours per week. And again, those were just two of the modules and there's much more AI can do. So Fabian, my my next question is in regards to the concerns. Okay.
what are the typical concerns that you face when you know in discussion with with clients and another question I'll maybe have two more if that's okay um how fast can I see success because we are being asked that quite often and also another question that is commonly asked um how does it look like in physical stores so as you said in the beginning pricing is a journey so therefore you you will not do everything at the same time. However, we can be very fast in kickstarting the journey. So, in our um typical implementations, it takes 12 to 16 weeks until we are live and can set the first prices. And then what what you will expect in the first price runs of course there's a lot of change because many improvements taking place leveraging AI probably for the first time.
So we typically see the first repricing being a bigger exercise requiring not only the the transformational change uh management efforts to make people trust in the new prices but also physically changing more prices than in an ongoing state. So that's uh I would say the the starting point and then going forward the amount of price changes deg dec decreases over time typically. Second h your question about the physical stores. So here we say the logic the theory is the same the models are the same to apply the rows are the same however usually the frequency of price changes is lower depending of course on the manual effort of printing price tags if there are no electronic shave labels in the stores yet. But in general the the dynamics are the same. There's KVIS sales drivers and profit generators. The optimization goals are the same. It's just a different scale and speed.
And how does this differ from by category? Sorry. So also here I would say the logic and the optimization goals um of the models are equal probably it's just the the extent of um price optimization and price change frequency. Thinking for thinking about consumer electronics for example where we would change or need to change the price multiple times per day for the KBIS this can look very different in grocery or fashion especially fashion thinking about the weekly markdown cycles towards the end of the season. Um so I think the frequency is a is a big factor. Um and second specifically for fashion which is a bit of a unique type of industry with this markdown um and the season seasonal cyclical um sales cycles here. Um this clearance automation is one of the key modules as we showed also in the beginning with the three lenses that uh is driving value. Thank you Fabian.
And you know I think that a question that is on everybody's mind and Jonah will be happy if you can address this. Is this a huge transformation project how does it look like? The easy answer to that and and that is no. Um so there are implementations uh that vary depending on customer landscapes and appetite to leverage the data. Um and that is both upstream and downstream within the wider organization. But in in real terms, it's not as big a transformation as everyone seems to think and that is part of the challenge today and that is we wanted to demystify pricing in that respect. So it also depends on complexity drivers. Um people believe that their skew count and the multiple channels that they have are big influences around the size and scale of this transformation but actually they don't add huge levels of complexity at all.
So business complexity drivers tend to come more like from the the retailers themselves with things such as legacy ways of working, their desire to sort of over rotate on on competitive price comparisons. Um but as you can see from from the image there, there are um upstream and downstream impacts where you can really leverage the data that is coming out of the pricing optimization engine.
So upstream for example you could look at demand forecasting and downstream you could look at uh rebates modeling as an example but people don't always appreciate that you can do so much with the data that's coming out so implementation itself um Fabian talked about 12 to 16 weeks we have genuinely seen implementations take more than six months but that's um that's that's the smaller um aspect we've seen some implementations take as short as eight weeks We generally start with a small pilot category and then prove value before roll out. But I think as a final point, not every organization believes in the power of change management. But in pricing programs, it's really really critical. Um so we really double click on that point. Pricing touches every single part of the organization. Therefore, every single part of the organization needs to be engaged uh for that.
We're firm believers in in change management. Thank you John and and you know this leads me to my next question and it touches the tension between IT and business. Uh there is always a tension. It's very it's it can be a positive one. It can also be a challenge. And Sh I want to ask you as an organization do I need to have the data in a perfect shape? So does the data infrastructure needs to be ready and perfect to start a project like this? That's a great question, man. So, that's the one that often comes up in retail transformations. Uh, the short answer is no. So, you don't have to need a perfect data infrastructure in place for embedding uh AI power pricing engine. However, you need to look both from the data strategy perspective that balances both business agility and IT scalability.
So the reality in retail is that the data is always fragmented and waiting for a perfect data infrastructure could mean you know you have lost opportunities. So looking at this API model price and recommendations pull insights from operational system. So you can see the different operational systems you know it is connected to such as competitive intelligence and product behavior analytics all integrated via API and SFTP and these fuel the AI powered pricing engine driving realtime strategic uh strategies and promotions. So instead of waiting for a fully mature data lake or data warehouses, businesses should start small, iterate over time and then scale. By leveraging these APIs for integration and AI models that have proven over time, uh it can actually lower it and business friction and they can also align ensuring agility and faster time to value.
and and Shri can you touch on the technical complexity drivers? Yeah, mo most of the times as John was already mentioning so it's the complexity will not be from number of SKUs and number of store variances etc. The technical complexity is more about how your data sources are organized uh what's your real-time data decisioning is and what are the any regulatory compliancy uh comes together. So for example in UK retail landscape a pricing engine must have or process dynamic VAT. Uh it also has to support multiurrency pricing regional promotions and any compliance uh you know issues. So all while integrating with all your legacy systems your ERPs your e-commerce loyalty market external systems. So all these factors will actually influence the complexity. So it's never about only the SKUs, stores or the variances.
Having said that, uh the modern pricing engines such as this are APIdriven and they are you know built and automated to handle high skew volumes, high store variances and high store networks efficiently. Sure. You know, we are being asked a lot about security and privacy, which often leads to the classic build versus buy dilemma that many organizations face. Could you share your insights on this topic? Uh security and privacy concerns are always, you know, critical for retail as well because given the strict compliance and regulation uh around the data transparency requirements. Uh however build versus buy is not a binary decision.
The reason is when you want to build something it comes with a higher cost and complexity and your speed to market will be lower and how can you continually update and monitor the changes when you're building inhouse uh will not necessarily meet something that is powered over time such as this. So security and privacy isn't just about buy versus build. Uh when you have something like buy option, you don't have to worry about your compliance certifications because that is taken care. You don't have to worry about your how do you handle your PIIs, how do you handle your end to-end encryption, how do you handle your multi-tenency. So all these are taken care in in case of you know buy options and another option is because this is given a SAS solution this can be deployed onto a customer's chosen cloud tenant.
So it need not the data need not leave customers tenant which means all the sensitive data that can be kept in retailers infrastructure and all the AI powered tracing engine can be still be adopted. Thank you Shri. And uh for my next question, John, um what if retailers are thinking that you know accessibility now to tools like Copilot, Gemini, OpenAI and and more that they can simply build their own AI pricing platforms because there are so many options out there. It looks very attractive. Um what do you think? Super topical question May and there are two main areas to consider. So Shre's already touched on it. One being buy versus build. Um my personal view and our view is really that um this is this is not a huge build product. This is a byproduct from experts and people with experience in this in this area. I would treat it in the same way that you treat a CRM.
Um invariably when a retailer is of size and scale, you are going to market to buy um to buy a CRM solution. our true pricing the same way. The second point is really that you know US and and Quick Lizard between us we've got a history of pricing we've worked for years creating refining um the pricing platform with the world's some of the world's you know best equipped minds those AI pricing models that you've just mentioned not those specifically but there are AI pricing models that sit within the platform constantly being trained on huge amounts of data um and the second reason you know that's that's why I wouldn't lean on the more widely available LM fulfill this requirement. It's embedded into the platform itself uh super refined and um yeah that's the reason that I would definitely approach this in that way and not leverage some of the other AI capabilities like you asked.
Thank you. Um you know I hope that uh we hopefully we've uh helped our audience to demystify some uh myths around uh pricing and leveraging AI. And for my last question, um I would like you Fabian to address it as well as you John, but maybe we'll start with you Fabian. Um you know the million-dollar question that uh is coming from clients is usually what is the best way to get started? I think as you say the most important point is to get started. I like the analogy of um permanent prototyping. So most of our clients um whatever we implement uh in the beginning is not what they end up in a couple of years of we start mirroring the asis pricing strategy and logic and then over time evolving it and stepping up in the maturity stages adding more and more AI modules. So just get started, the rest will follow. And Fabian, I absolutely agree.
Facing into pricing optimization, it's not something to be fearful of. I want this audience to to really take that as a takeaway from the session. It's actually the opposite. It provides one of the most powerful and accessible levers to to drive performance. And there are two main routes to get started. I'd suggest one is probably reaching out to us via the contact details page that's embedded on the last slide and in parallel I'm sure mine is about to tell us but there is a pricing maturity questionnaire which I would encourage everyone to take on the back of this session today so indeed John uh so to wrap uh up the webinar we have something actionable and valuable for you a quick pricing maturity assessment so within just a few minutes.
Um, once you complete the questionnaire, you'll receive a personalized assessment delivered straight to your email box uh helping you to understand where you stand and what you can optimize next. It's a great tool to get started um even as as an internal um process to understand where you're currently at. So, please scan uh the QR code that you see here. Also, please feel free to reach out to us either to John or Fabian if you have any further questions. Um, you know, I want to thank you. Um, I want to thank you for joining today's webinar and pricing management. Uh, we hope that you find the insights, strategies and real world applications valuable for you uh to when you navigate the complexities of pricing uh in an ever evolving market. A special thanks uh to Fabian, John, Shri. Thank you so much uh to our experts for joining today and sharing from their experience.
Uh thanks again for being a part of this session and uh have a great day. Thank you so much. Thank you. Thanks. Bye-bye. Bye.