PegaWorld iNspire 2023: How Vodafone Uses Pega to Deliver a Value-driven Retention Experience in Assisted and Digital Channels

There are some questions that are consistent across all telco upgrade teams: How do you drive the best upgrade experience? How do you ensure you’re not losing the customer to a competitor? How can you drive upsell? How do you reduce dilution?

Discover how Vodafone is using the Pega Customer Decision Hub™ and Pega Next Best Action Advisor to deliver the right offer for the right customer – to ensure the customer has an easy decision, lower fall-out rates, reduced churn, improved post-upgrade value, and improved experience, across call centers, retail (using tablets), digital, app, and other touchpoints.

Transcript:

- Welcome everybody, to the Vodafone presentation given by Vodafone UK. My name is Axel Wells. I'm a industry principal here at Pega, focused on the communications vertical, and I am really excited about this presentation. I had the opportunity to work with Sameer and Russ who I will introduce here in a second. Over, I think it was two years ago at PegaWorld 2021 when it was virtual, and we joked, hey, we're gonna come back and we're gonna tell the rest of the story in the next PegaWorld. Well, that has turned into 2023 and I'm really excited about this story that they're gonna tell you about how they have rolled out their always on marketing implementation. It's gonna be an exciting presentation and today you're going to have Russ Welton who is the decisioning authority lead, Efe Asci, who unfortunately will not be here in person, however you will see Efe present. And Sameer Prakash who is from Adqura. So with that, I'm gonna hand off to Russ and I'm really excited about this, Russ.

- [Russ] Yeah, at least one of us are. Hello, everyone. Nice to meet you all. Really glad to be doing this in person. just biting the bullet and doing it in person is gonna make a real difference. But for those that don't know who Vodafone are, so Vodafone is one of the leading UK and UK and Europe and Africa telcos with a customer base of about 18 million. Our network coverage is about 99%, so we're currently in the migration from our 3G technology, turning that off and moving into 4G, and really expanding what we do in 5G. One of a recent use case of that is the King's coronation just recently, it's first ever broadcast of that over a 5G standalone network. Something that we at Vodafone are really, really proud of. To go into what we're gonna talk about today, so we, Vodafone started on this journey about four years ago and we kicked off what we see as a move from product centric to a truly customer one-to-one engagement cycle. There's a lot of detail on that slide, but I think we have the best way of showing it, which is a nice two minute video and I'll let that take over.

- [Voice Over] Every Vodafone customer is unique, in their interests, in their behaviors, in the conversations that really resonate with them. So it's time that we move away from a one size fits all approach and a fragmented customer journey, towards a journey that's always connected and tailored to the individual, all driven by the extraordinary power of always on marketing. Using our collective potential, we'll build a new customer experience that responds to them through personalized one-to-one communications. Together with real-time information and analytical insight, we'll be able to deliver highly relevant messages to each customer, based on what we know about them and the places they've been. As well as greater relevance, we can ensure everyone gets a truly consistent experience across any channel at any time. And gradually, through adaptive learning, we'll better understand our customers with every interaction. So what does AOM look like through a customer's eyes? Imagine someone has just landed in the USA, they always use social media on holiday, which has meant big roaming charges in the past. AOM decides in real time to serve them a Facebook ad for our unlimited max plan, with roaming in the USA included. They don't take us up on the offer straight away, so later when they browse online, they see a banner with the earlier offer. This time they do take us up on it. Then when they return home they get a text from us welcoming them back to the UK. This is the future of our customer experience. The future is exciting. Ready?

- [Russ] So I really enjoyed that video. I remember us all briefing that together, it was really, really exciting. The soundtrack to that really resonates with me. I thought it was really cool. I think the problem statement that most of us in this room are all coming here or on that journey or already implementing, is that we are coming off of systems that are 10, 15 years old. We had the original accordion implementation 10 years out of life support and we're also on Unica. That basically meant that we have no common personalization and where it was basically resulting in really poor customer experience and missed value opportunities, something that we drastically needed to improve on. So when we set out for this, we set our mission. That mission was to really transform how we communicate with our customers in both inbound and outbound, and we'd done that for a series of value drivers and customer experience principles. One we're mostly talking about today is definitely the upgrade value but all resonate in every single customer touch point that we have. That then brought us onto the solution. The solution that we set out to deliver was always on. It was bringing in realtime customer data, big data, AI and adaptive in every single customer touchpoint that we could deliver. I'd like to hand over, he's gonna talk in a bit more detail to Efe on how our upgrades journey actually works.

- Hi, everyone. I'm Efe Asci. I work in Vodafone customer value management team as a senior base growth marketing manager. I'm sorry that I can't be present in the conference room and meet you all today, however, in this video presentation, I will walk through our upgrade journey principles, how we use always on marketing to meet customer needs and business objectives and our value driven retention recommendation logic. Upgrade journey is one of the most important journeys in the lifecycle of our existing customers, their contract through a new handset or offer. To shape this journey, we have identified customer needs and designed our four step journey with these principles. First one is where we analyze customer needs, based on all customer and interaction data, and we prime and ready the customer for the upgrade with our recommendations. This is where we collect all the data and process it to understand customer's needs. The second one is the recommendation. Based on what we know about the customer, we guide the customers with our best offers, as well as let them retain the control, by allowing them to assess and compare the recommended devices and plans. Before the customer decides at the third step, we make sure that they feel in control. Customers are not restricted with the recommended offers. Both in digital and assisted channels, it's very easy to go beyond their recommendations. We get the feedback from the customers and amend the offers if they're not happy with what's presented. Lastly, before completing the order, we also recommend relevant add-ons such as international bundles, insurance products, or any other relevant add-ons. So this way, we aim to create a simple upgrade journey where we address different customer needs at every step. In addition to the customer expectations, I want to also highlight the commercial objectives here as you basically address both customer needs and business objectives in our always on marketing solution. First objective is to ensure that we have the right competitive offer set at the individual level. There's no pricing set that fits for all customers in the world of one-to-one engagement. Therefore, our always on marketing solution allows us to tailor offers at individual level, driven by churn model and our segment level pricing. This is an important part of our logic, as we rerun our recommendations based on customer interactions or allow agents to check other options as well. Second objective is to bring up the best offers from the available offer set and we do this through our recommendation logic. The recommendation logic achieves meeting customer needs and business objectives, picking the best offers for value and conversion using the propensity models and business levers. Last but not the least is the customer experience across channels. Both digital and assisted UI, designed to have a very simple experience and flexibility to compare the plans recommended. In this example, customer sees their current plan device and monthly fee, together with the recommended device, recommended plan, their total monthly fee and the personalized discount. In this slide, I will zoom in on our personalized offer strategies, which we group as grow, retain, and save. This chart shows the direction of different offer strategies and the center of this chart represents the customer's current state. Grow refers to the section one where the customer gets a bigger bundle with a higher price. This is actually an opposite offer. Saves offers generally sit in section two and three with similar or bigger bundles at a lower price, targeting mainly the high churn risk customers. And lastly, retain is actually referring to the similar bundle and price point compared to the customer's current bundle. For each customer, we have three recommendations, driven by models and business roles, and it could be any combination of these three strategies. For example, number one recommendation could be a grow and number two and three could be retain offers. So with these personalized offer strategies, we decide the direction and magnitude of the arrow from the current state for each recommendation position. In the next slide I will explain where our value driven strategies sit in overall recommendation logic. In this slide I will explain how our top three recommendations are shaped. So the top three personalized recommendations are shaped by the big data models and commercial levers. I will explain how model driven inputs and commercial inputs are used together to prepare the top three recommendations. Let's start with the churn model. So based on the customer data and interaction history, we score each customer's churn propensity together with churn reasons. This output is used to influence both customer's upgrade strategy and the pricing offer set. So the first area it's linked to is actually the pricing. Our pricing set has multiple layers in addition to the standard offers. The first layer is driven by churn model that enhances the offers and discounts for high churn risk customers. The second layer is segment level pricing. That increases the number of offers for predefined customer segments by commercial teams. So actually, commercial teams have the flexibility to open up extra discounts or offers for certain customer cohorts. The third layer is the channel toolkit, that unlocks extra discounting at channel or team level. This is basically some teams in telesales that mainly have the sales conversations. So all these offers and discounts are then used by our big data upgrade propensity models. You see them in the third section. So our upgrade model decides three things, the likely upgrade path, which is or device, the propensity scores for each device and the propensity scores for each plan, and rank them from the highest score. So just to highlight, so far it's purely based on the propensity to upgrade, ignoring any commercial target. This is actually the fourth area. The fourth step is where we include the commercial levers, personalized customer strategies and upgrade revenue targets are defined at this stage, again, driven by churn scores and segment levels or strategies. So if you see the arrow from the churn to commercial levers, basically, customers personalized upgrade strategy is defined by all churn score and again, with commercial levers, defined by the commercial teams. So output of all the commercial targets are then used to pick the plans with the highest propensity scores that meet the commercial targets. This way we come up with three relevant recommendations that meet the customer needs and the business objectives. These offers are presented across digital and assisted channels and rerun with any change in customer's data. I hope this has given a good overview on how we use both big data models and commercial levers for value driven recommendations.

- [Sameer] Okay, thank you, everyone. Hope that was explanatory. If you have any questions, you can come and speak to Russ and myself after our session. I'll cover the next section. I want to talk about the experience. So before I show that to you, we used next first action advisor, which is a Pega product for delivering the experience in the contact center in stores and we have built a headless service that we are using in the digital channels. So the steps on the top are the steps that Efe was talking about, preparation, recommendation, negotiation and upsell and cross sell. And within that at the bottom you can see the snapshots of what our user experience looks like. I'll walk you through the contact center ones and then cover a bit about the web and mobile channels after that. So this is our contact center and a retail front end that we have. Just before we step through the details of what we have, on the left hand side you will see our legacy system, which incidentally was Accordion, accordion, recommendation advisor. And a point to note was it is a very e-commerce type of solution where you are picking up things and dropping them in the basket in negotiation with the customer and there was a budget that you were targeting on the right hand side. We have done away with that approach. The approach that you see on the right hand side, which is driven by CDH, is effectively three plans bundles that we have. You can see three of them vertically at the bottom, and they're pre-calculated for every customer. So it's a recommendation driven approach instead of putting your bundle together in discussion with the advisor. Within that screen itself, if I expand that a bit more at the top, we have what we call profile and usage. We will show you the profile a bit more in a minute. But effectively, that's giving the advisor visibility of a summary of how the customer's using their plans, their handsets. At the bottom are the three recommendations. And obviously, the first one that we have on the left is the top recommendation, and you'll see there are some call outs, for example, there's a purple promotion tag on top, et cetera, which tells the advisor many things. There's lots of icons in there. One of the icons is they're part of a Vodafone together plan. That means they have multiple lines which are shared, they're getting discounts, et cetera. You will also see, let me just point it out, you'll see there are certain colors that we are showing in green and red. And again, they're indicative of if the agent or the advisor were to close these deals, does it meet the KPIs, does it meet the targets for them for that particular deal? Lastly, on this particular screen, on that, you will see there's a refine by panel and with the refine by panel what we are doing is we are allowing in negotiation with the customer to specify if there are specific preferences they have, for a Samsung handset instance, they're looking for a 5G handset, and whilst we have the recommendations to go at the point of launching the front end, what this allows the advisor to do is then tell the CDH in the backend, this customer prefers Samsung handsets, so we will refresh and return the recommendations which are in line with what the customer's preferences are. In terms of the bottom right, you will see there is additional hidden details in there which explain why we are making certain recommendations. Again, the purpose of this is when you are actually to talking to a customer, you'd say, the advisor says, I think this is valid for you, why it's relevant. Again, we are using this information. It's personalized for the customer, it's based on how they use their devices, it's got language that marketing and comms team have put together to explain to the advisor what to say to the customer, what things to call out while having those negotiations. So it's extremely contextual, it's driven by rules and AI that we have in the backend. Just expanding that screen a bit more, on the top left you will see, if you click on Joe Bloggs, it creates a modal popup, you can see more details about Joe Bloggs, you can see their dress, you can see the different lines they have. You'll also see on the top right it's got some information. So in this example for Joe, they've requested for a put out code. So that's highlighted. Any event that has happened recently in other channels, we are able to show this as well. Some of this is down to what Russ was referencing earlier and we will talk about it in the architecture slides. We are getting interactions from all channels across the Vodafone estate. So we are able to ingest those events. We are able to see the relevance of that and then present that processed information through CVH onto this UI for the advisors to use as part of their discussions with the customer. Now, when you look at the three recommendations, at the bottom, I'll point them out for you, the three that we have here on both sides, these top three recommendations, you might say, I don't know, I'll take Russ as an example, Russ might say, yes, it's great you have an iPhone, but actually, I'm not interested in an iPhone or this iPhone is too expensive for me, I'm looking for a less monthly payment for the device that I have. So the approach that we have is, let's start with three standard recommendations, but if you click on that pencil icon, it gives you a popup where you are able to refine. So you can change things like you can change your device, you can change the plan, you can change the add-ons, et cetera, and advisor is able to click through these and you'll see there's certain filters on top, you see the ability to check for stocks, you'll see the device price, et cetera. So what we are doing is we are giving a number of features within the front end for the advisor to interact and build the perfect bundle for the customer so that we can retain them with, okay? In terms of what next, once you have selected, you have refined, once you have finalized what the bundle looks like, the advisor would scroll to the bottom, they select the bundle which has been picked up and they see a basket summary screen. Within the basket summary screen, there's sections that you expand, you can talk to the customer about the additional benefits they're getting with what they have selected, the discounts which have been applied, et cetera. And at the very bottom you will see we also have the ability to, well, what we do do is we will remind the advisor a next best action. So this is the fourth step that Efe was talking about, ad insurance. So this is the place where we are able to say, there are some additional add-ons that you might be interested in. If the customer says yes, we click on it, and then the system returns what those additional next best actions would be. Lastly, we click on the continue screen. No sorry, button at the bottom of the screen, which triggers single click fulfillment. I'll let Russ talk about that when he covers his architecture slides, how that works. But in essence, the UI that you're seeing is making a loads and loads of real time calls, it's assembling all that data, making a recommendation by parsing it through the AI and rules engine, and when the customer's happy, you click once and that's it. The complete fulfillment in the backend is taken care of by Pega. Let me show you the equivalent experience that we have in the digital channel. We've been live in the contact centers, we did the pilot last year, we started before the iPhone launch, then we had to wait for iPhone Black Friday, Christmas, and then we did the rollout from end of January. The digital experience is ready, it's currently in friendly user testing. We're expecting to go live with that in July. And you will see on the bottom left is a representative email which has been sent to the customer. It's very personalized. It says meet your new phone, et cetera. And when you click through on it, you drop into the upgrade journey, which shows at the very top, you will see the recommended handset at the bottom, it shows what they currently have, and everywhere you see those red call outs are next best actions being returned to the front end by CDH in the backend. So we return the reasons why this handset's relevant, what is the top bundle that we have? We show any personalized discounts which are available for that customer. If the customer wants to look at alternate handsets and they say, let me show other handsets, We do things like we calculate what's the trade-in value for this handset. We offer higher trade-in values if the customer is a higher value customer. You will see on the right hand side, there are different categories of handsets we show. So for example, we've got a handset which is good at taking photos. We've got a handset with a bigger screen. So we are recommending top categories for that customer. So it's personalized. And within each category, we're returning the top three handsets that we think they'll be interested in. Now, it's possible that the customer's not interested in the handsets we are showing, they can scroll down and look at the complete list of handsets that Vodafone sells. All of this is things that we learn from and from an adaptive perspective, what was the customer interested in? Did they like the things that we showed them? Yes, we learned from it, no, we learned from it, and then we refine it the next time we have a customer on the website. Let me show you the equivalent mobile journey on the right hand side. So just a stat for you, nearly 50% of our upgrade journeys are currently happening in digital, more than 50 now, in Vodafone UK. And nearly 70% of digital upgrades are happening in the mobile channel. So if you can imagine on that small device that you have, being able to scroll through handsets and handsets and plans and plans is not feasible. So you can see it's a very recommendation driven experience. You can see more phones, main reasons why, et cetera. It's still using the same backend services but it's optimized for a mobile experience that we are delivering. Okay? Now, how does it all work? I'll hand over over to you, Russ.

- [Russ] Cool. So this is where, actually, I feel a bit more confident. So we set out on this journey, we've already implemented CDH solution. We were fully live, fully onboarded with Pega CDH and our MBA and commercial teams were fully operationalizing that. So when we looked at from an upgrades perspective what tool did we want to use, how were we going to integrate, Pega next best action advisor was the obvious choice. It delivered the capability that we were looking for, it provided the architecture we were looking for for based interface that we could then reuse in multiple different channels. Our whole experience through Vader is delivered for a mashup using the Pega mashup capabilities directly into our CRM system, and our agent is completely able to utilize that system along with Vader to come up with the best deal for that customer.

- [Sameer] So Russ, just to call out, Vader is the name you've given--

- [Russ] Sorry, I'm not allowed to. VADR is the appropriate name, I apologize. Vader is the name that we've given as an application but it is next best action advisor at the heart. So obviously I talked a lot about restful interfaces. We are fully rest based. We looked at making reusable journeys or reusable APIs that could be used in both digital and assisted. So we set out with an assisted journey that would deliver against those four key pillar journeys that would be associated with that journey that we talked about with FA and submit. So we first start with prepare where we have a series of different APIs that will allow us to get our customer details, get those recent events, and also get things like customer usage. That's fully delivered by CDH implementation, but we also have a number of three or four different external APIs that we also use to power that front end. Now, as that agent negotiates and has that conversation with that customer, we have a powerful recommendation API, both the MBA and the refine space that talks to our strategy logic, talks to our data that we've gathered in the front through the previous stage and working with our MB execution team to be able to establish what those offers will be in control and have that level of control that they will need. Now, as they go through that negotiation phase, it's really chatty. Every single interaction, that agent is talking to that customer, working out a new deal, understanding what they're gonna do and those services need to be really powerful to be able to deliver that and also fast to deliver and fast to bring back. And Pega's allowed us to do that through three series of APIs of product list, validating the basket and validating any deletions. That validation function is powerful, it's continuously learning based on what that agent is doing with that customer and always retaining those best plans based on what they've chosen. Finally, we go into the submission phase. That submission phase, like someone said, is a really, really complicated API. It's a really complicated journey. Lots of processes, lots of backend processes. So as a telco, and I'm sure I've got a few telcos here, we have lots of standardized products that need to be added into that product, but we don't show our customers because it's like, 4G, 5G add-ons, they're expected, they're not needed to be shown on the UI. So that all is orchestrated through the backend and handed off into our CRM system where that agent can continue that journey. I wanted to talk about, you've heard a lot today about customer 360 views. We aren't the same. I talked about it or we talked about this in 2021 remotely, like Axel said. I wanted to touch on it again. The one main difference is, is that in an agent channel you have to be as close to real time as humanly possible. That agent is changing things on the CM system before they launch into us, they're changing things inside the time they're with us and we have to be able to react to that. And what we've done with all of our customer touch points, all brought into what we call the Customer XCAR, all through a series of web services, API interfaces, realtime streaming and collected together through a series of Pega data flows. It's really powerful, really cool tools and what we're looking at in coming releases is looking at what CPD will enable us to allow business teams to get better data out of the raw data itself. The strategy framework, it's not really my forte but I wanted to show mainly focus on two points. It can't be fully automated. You need the control, you need your MB execution team able to manage the eligibility rules for that customer upgrade. And the way we've done that is providing certain portal, certain capabilities that allow them to do that. And then further on the far right you'll see that our commercial teams have a UI based access to control what we present in our recommendations, both through commercial recommendations, managing customer segments and also things like tenure and understanding what is the best for our customer. That control is necessary to make sure this experience is worthwhile. Now, everything that's underpinned here is done by the product catalog. As a telco we have thousands of products that continuously turn, continuously change in terms of eligibility and we need a solution that would enable us to have that externally managed through our CRM system and provided to us in our CDH base. The way we've managed to do that is we take about 28 feeds, I think it is, that come on on a weekly basis. We update that into our Pega database and then we orchestrate that access outwards for a series of activities, a centralized cache and then access through our strategy logic through the data page implementation. What that's enabled us to do, really, is go, all that complex logical processing of what's compatible, what's sellable, can be done easily and accessed by our logic team, that access that data. Finally, our MBA execution team, a product catalog is probably not enough on its own. You need to be able to control commercial aspects with that such as sellability and maybe some things like what names you want to give it, those characteristics that you see on UI, you want that ability to do so and we've enabled that business team to do that through our business portal. Now obviously, I don't how many of you are technical or what next, but I wanted to make sure we had time for going over some results. Unfortunately, Efe isn't here, so you're gonna have to deal with me doing this. But we are now fully rolled out to our telesales. That means we have four sites, we have two that Vodafone are internally managed and two that are externally managed. We've managed to roll that out. It was a really, really long journey over about a year where we've done lots of different pilots, rolled out different types of agents and make sure we captured that feedback. And as we rolled out, the whole team went there, I was there supporting on the floor, making sure that went well. And some of the feedback you can see on the screen is the feedback that we got was one of the best implementations that they've seen in a long time. I'm really proud of that because I was central to it, Sameer was central to it and a lot of the people in this room, I won't call out by name, were also really privy to it as well. Now, the actual results of this, I can only present a few of them, we're still going through a lot of the results and trying to catch them up. One thing I'd really like to call out is the average handling time. So with a recommendation based approach where you can present our recommendations thing, we managed to decrease that by 19% from our legacy systems. We managed to increase conversion and increase our average net revenue by five percent. It's really great results and looking forward to what we can do, as we roll out more and more capability into this space.

- [Sameer] Can I just give a voiceover, just one more on that?

- [Russ] Yeah.

- [Sameer] So in terms of the average handling time, the main reason it's come down is in our legacy system, as I was showing you in the screenshots, we actually put the basket together while talking to the customer. But with Pega, we get a recommended set to start with. So we're saving about five minutes per call in the upgrade journey and we start with the recommended set, we are refining it but we are learning every time that the customer's asking to refine it. So we are expecting it to get even better. The second thing to call out is in UK, in Europe, we've got a standard of living crisis, so the prices are going up, it's really hard economically. And we can see that. So for example, handset attachment, insurance attachment, et cetera, are down. However, we are still doing five percent A and R. We did a comparison of January versus April. So the numbers are up, we are making more margins, the value of the basket is higher because the customers are reflecting, understanding that the recommendations we are making are linked to how they use their products and services and hence the acceptance is higher than what it was in the legacy system.

- One thing I also want to say is make sure we took away some takeaways and considerations, first of which is performance. Like obviously you're facing into agents with this type of solution, you need to make sure that you cannot be... Your UI is fastly readable, especially in the digital channel as well. We've taken lots of strives and lots of continual improvement, like there's another takeaway to make sure that we improve over time, but also making sure that you're customer centric in what you do. Thinking not only as your end user customer, making sure my recommendation's perfect, customers as well in this space, they need to be able to have the tools that they need to make sure that they can do and operate what they do best, and that is negotiation. Our continual improvement journey really started from when we get going. Every single , every single time we had a new capability, we were getting feedback from agents, making sure the UI is something that we wanted, they were happy with, they were happy to use, and they delivered their key requirements, and that hasn't stopped from the initial delivery onto now, we're continually iterating, continually delivering new capabilities and continually improving the core product set and performance that we have. Now, let me hand it back over to Sameer. He's gonna talk about what we're going to be doing next.

- Yes. So that's the last slide we have. So are we finished? The short answer is no. So what are we doing? Small letters. I'll talk you through that. The first one is commercial recommendations. So as you saw and Efe was talking about it, we're trying to bring the commercial teams closer to the decisions we are making. It's linked with the third point, which is around simulation. So what we are now doing is the commercial teams, they set their targets, they refine every two weeks, every fortnight, so how do you feed that into the system? How do you run different simulations? So don't just set it up, try the what ifs, see the boundaries, define new cohorts, define new recommendations, and then that gets implemented. So that's one. And the second one is adaptive. We're trying to push out adaptive more across the board, not just from handset recommendations, but understanding from the business what are the outcomes they're looking for. Now, that's really important. I'll just take a few seconds. When you define an adaptive model, you don't just switch it on, you speak to the business and find out what's good. So that's the positive. And also what's bad. In most cases, it's not that black and white, especially in a commercial setup, so we have to agree what inferred positive and inferred negatives look like. And if you get that right, then your adaptive models will excel. So there's a lot of work happening right now for us to get those adaptive models at that level that we want. That's it. Back to you, Axel.

- [Axel] Any questions? Well, thank you, Russ and Sameer. That was a fantastic presentation. We will open it up to questions now. If you have a question, we have a couple of mics up and be happy to answer the questions.

- [Audience Member] Pull it down. Quick question. Great presentation. Related to the customer data platform, are you capturing all of the customer's clicks throughout the website and the app and then using that to feed into next best action? And are you streaming this or are you caching some and doing nightly feeds? We've noticed that the sheer amount of data that we're trying to stream in real time is causing latency.

- Yeah. So we are fully integrated with... So we have a tedium, realtime tedium stream that tags . We use tedium in the website that tags every single event, every single web click. That is streamed in real time to us in CDH and we use Pega real time data flows and Event Strategy Manager to capture that data and aggregate it upwards to present back into our CDH. I think latencies, at most a second, in most cases, but we haven't really foreseen so many issues in that space, but that's how we're currently integrated and we're looking at using a CPD, or customer profile designer, sorry, I've got to remember acronyms, to use and make sure that we can aggregate that data in a more effective way for our business users.

- All right, any other questions? Yes.

- Thank you and really good presentation. What's your trigger point for the contact centers reaching out and having that conversation with the customer? Is it purely reactive when they approach you to discuss an upgrade or do you do anything proactive or is there even anything, if they have a service issue, do you then filter the conversation into a conversation about upgrade?

- Yeah, so from the journey that we just went through, it is purely reactive but that is also tied up with a lot of outbound communications and outbound management. We also manage through CDH through proactive campaigns or through a proactive MBA. Sorry, I've got to be really careful what I say. my business. But we have a continual like... So we have both real time triggers and a generalized batch of decisioning every day that we run that allows us to target customers and we'll outbound call them if we deem it necessary and they're in that relevant space. Obviously we try to move them more towards a digital experience first. And depending on the cohort of customers, obviously you do see a lot of people going online, look online, look at the deals and then call into the call center with a deal in mind. We're looking at capabilities to make sure that we can see that or take that transition from the digital experience down into the retail so it's much more seamless for our customers but it's something that we're looking to do in the future. But yeah, we definitely do have a reactive and proactive campaigns in this space.

- [Sameer] So yeah, if I can add to that, in the example I was showing you in the digital journey, so it starts from an email. So as Russ said, we send them to the digital state first. So that's one. We integrated into the IVR channel as well. So we get voice to text translation by nuance. In real time they pass us that information and we use that to make proactive offers to the customers. If we think we need to have a retention conversation, we'll give them the link which will then drop them onto the website via SMS. And lastly, again, as you were saying, Russ, we got realtime triggers. For example, if you submit a port out request, so in realtime we calculate, do we need to contact this customer, are they at the end of their contract cycle? We do it for data exhaustion events, for instance, and then we will proactively call them from a contact center. So then we use the same UI.

- Okay, thank you.

- All right, any other questions out there?

- [Audience Member Two] Just one question from me. How did you get the commercial teams engaged?

- So at the start of our engagement, we did a six week piece of work which was sponsored by the chief commercial officer and we got his direct reports to come in. We looked at what are the current journeys across all the different journeys that they have, what are the different moments of truth? We used the data call data to find out where people are dropping off, where the opportunities are. And then through a series of workshops, we identified, typically, what are the highest value moments of truth, which then led us to create a roadmap of engagement, which was our engagement plan for the always on marketing roadmap. Now, as part of that exercise, we engaged with the commercial teams, we engaged with the other teams as well. The second thing to call out, we won Greenfield. We already had Accordion recommendation advisor, which is the previous iteration of what Pega bought out, Accordion. So they already had commercial teams who used to operate within Accordion, so they were aware of what's involved, et cetera. Obviously now we moved them up to Pega's way of thinking.

- Great. Any other questions out there? All right, I think we're gonna wrap it up. I want to give a round of applause to Russ and Sameer.