[Music] hello everyone my name is is MF I'm a senior specialist Solutions architect for data analytics at AWS and today I'm going to be talking to you about maximizing business value from data now let's talk about the aspiration of most organizations these days most organizations want to become datadriven organizations and everyone understands that data is important that dat is a treasure Trove that is untapped and waiting to surface additional business value for their stakeholders but how valua
ble is data really well looking at a study that Forbes conducted organizations that made data just 10% more accessible realized a $65 million increase in net income which corresponds to a 45% Roi over 5 years and a 48% uh reduction in total cost of operations in the same 5year period so the the business value to be gained here is massive from a general perspective because of this 85% of businesses want to be data driven but only 37% have been successful now Accenture calls this appropriately the
data value Gap the gap between actually realizing uh business value from data and the current situation of most organizations and in fact 68 68% of organizations reported that they're still unable to realize value from data and only 28% of respondents to a survey that Accenture had conducted have a data strategy in place so what does it take to actually close the data value Gap well it starts with having a modern data strategy and having that strategy corresponds to a plan and that plan will re
alize um into execution which will help you achieve that vision of becoming a data driven organizer Iz ation now one of the problems is actually knowing where to start because the one of the reasons there is a data value Gap is that significant effort and investment are required from a people and processes perspective to actually uh become a datadriven organization so to justify that effort it's important to understand what are the highest leverage activities I can do as an organization where am
I going to get my biggest bang for my buck in terms of data projects and to think about how we can uh rank different initiatives we start with the business outcomes that would underpin a modern data strategy starting with the ability to make better and faster decisions and improving customer experience and loyalty staying ahead of the competition and using data to to build a differentiated value proposition preparing for the future and future proofing our operations our plans our Revenue Invest
ments our cost optimization initiatives Etc and of course the cost optimization and operation uh and operation optimization initiatives themselves looking at these five business outcomes we can then translate those to the highest leverage um initiatives or activities um that you can undertake as an organization to get you started on your journey to become a datadriven organization So based on our experience working with Fortune 1,500 and Fortune 500 100 companies and all uh organizations of diff
erent shapes and sizes around the world um we've identified some key use cases that can maximize the business value for you if you make uh the right Investments um in the right parts of your organization and transform it in the way that is needed to realize such business value so the use cases we're going to be talking about today customer 360 is one followed by supply chain optimization datadriven decision- making intelligent op intelligent applications and cost optimization so let's start with
customer 360 this is um uh a word or a term that has been uh thrown around a lot quite lately so by our definition for the sake of our talk it is a unified and enriched customer profile purpose-built to power marketing and advertising so why is having customer 360 by that definition important well simply leads to increase Customer Loyalty improve marketing and augment sales but to do that you need to become a customer Centric organization you need to own the customer experience end to endend an
d that requires to H that requires you to have a couple of things having an omni Channel customer experience requires number one integration with all consumer touch points so the different parts of the customer Journey you need to be tapping into that and having uh data points corresponding to their experience in each part of that Journey the second point is taking those um data points and producing actionable insights uh from uh from them and the third is having unified customer data taking act
producing actionable insights from touch points by themselves isn't enough the customer um has a certain experience that they have by themselves with different parts of your business so to have the same view that the customer has your view needs to be unified the way that the customers is and that's that's why having a unified um picture of customer data is so important so to do these three things it requires you to be a different type of organization number one a customer Centric organization
as we said and a data Centric organization that creates this kind of mentality and process where you are continually continuously learning uh about your customer and optimizing your interactions with them so looking at this flywheel on the right you'll see that the consumer profile um then overlaps with analytics and decision-making engines that you would build uh inside your organization which would then feed into uh activation campaigns and other uh engagement vehicles with your customers whic
h then get fed back into the profile and the customer experience improves over and over again as you iterate through this cycle so obviously to achieve all of this knowing your customer Inside Out is job number one and that means knowing their demographics their psychographics the transactions they've made their purchase record records support cases and tickets that they've opened with you uh their shopping habits their content preferences and many many more now currently the state is that less
than 10% of companies have this kind of 360 degree view of the customer and even less than that use this view to systemically grow their business so to look at how how a customer 360 um solution whatever it may entail would be set up on the left you would have the different data points coming in that tracks your customer uh Journey or experience or engagement with you as an organization including product searches um their experience in your brick and mortar retail stores the devices that they're
logging on from whatever uh tickets they um whatever tickets they've logged with customer support and of course your backend operations in terms of how the customer interaction reflects in terms of inventory and orders uh and uh cart activity and engagement on your website etc etc you take that data and you unify it into that single customer profile which then can then be used to serve Downstream systems using things like marketing micro segments uh offering personalized offers to your customer
s looking for the next best action looking for different opportunities to upsell and cross sell looking for opportunities to prevent churn uh using aggregate data from different customer behaviors to forecast uh demand in the future looking for different patterns taking into account the effect of different pricing and promotions all of that is enabled by having that unified customer View and mapping it to your customer Journey now this sounds like a lot of work is it worth it the answer is yes b
ased on a study that apqc um conducted across 300 organizations who have undertaken a process of building uh customer 360 capabilities inside their organizations they've experienced an almost 44% reduction in sales cycle duration a 22.8% increase in customer lifetime value which is a key metric for measuring uh the the amount of Revenue that you can uh expect to experience from a customer across their lifetime with your product or service or organization at 25 .3% faster time to Market which mea
ns you can uh react to external signals faster than competitors can you can respond to your customer needs uh faster than they can and significantly a 19.1% improvment in NPS or net promoter score rating which is a crucial marketing kpi for indicating your uh customers's perception of your brand so across all of these key metrics um this is the this is the kind of value you can expect to realize if you undertake such an initiative so we talked about the business benefits of implementing customer
360 but how do we actually put that into action well the first thing we need to note is that customer 360 is not one thing it is a combination of uh combination of tools combination of platforms and Technology uh that work together in tandem to produce the business outcomes that we just discussed so we look at a combination of first Party Services from AWS and from the Amazon Cloud with partner solutions that are the best in class at what they do uh across different uh mtech and attech players
uh so that you can U have a combination of buy and build Solutions together that will produce a complete endtoend customer 360 view so starting at the bottom layer the data infrastructure and Technology layer now since you have all of these touch points that you need to interact with you're going to be uh you're going to be ingesting a lot of data across your call center your Erp your CRM um across social media data across chat and call data emails images video or what have you um so the technol
ogy platform ingesting this data needs to support structured unstructured and semi-structured data so that you can land it and have it in one place and then you can combine it with uh solutions from segment for example so that you can organize that data into to a customer 360 view now once you have that foundation in place you can start to produce actionable insights from it and that's where you go to the analytics and decision layers this is a layer that helps marketers make the right targeted
decisions continuously at scale to move customers across the journey quickly so not just one or two campaigns but continuously learning about customer behavior and leveraging machine learning to um figure out the next best action and use other data points to enrich the customer profile and have even more targeted uh better um uh next best actions that improve the customer experience and then finally we move up to the activation and experience layer and this is how customers are experiencing a br
and from advertising Media Mobile and and different types of um um touch points uh that all uh are the interface to your brand this is basically how your customer is interacting with your your organization and this is the layer where marketers design the customer journey in terms of awareness interest testing buying and hopefully turning the customer into um a loyal customer so combining all of these three layers together with a combination of uh AWS um Solutions as well as uh partners that we w
ork with you can then um have that end to-end customer 360 View and build that Unified customer profile that will hopefully improve your net promoter score and your cessat score for your customers and uh reduce churn improve Revenue um and any other relevant metrics uh like the ones that we talked about in the aqc report to put that into practice one of our customers Cox Automotive had a challenge they were concerned that the loss of third party cookie data impacted their ability to provide pers
onalized offers to their customers so what they did was they built a 360 view of households that can be utilized across many different business units and they did that from a technology perspective by building an identity graph that defines the identity and combines Shoppers or or prospects of a household and different leads that they could um explore um this is a very large graph it consists of approximately half a billion edges and uh4 billion vertices um running on a a graph data base and uh
the results were that they yielded twice as much browsing history per household compared to using individual cookies and they were able to increase their average online interactions per consumer by 180% a 380 per increase in average cookies per customer and they have six business use cases that benefit from the work they've already done to benef to create that identity graph so it is a 1 + 1 equals 3 type of situation uh even if you don't know in advance all of the use cases that customer 360 wi
ll enable even the immediate business value is enough to justify the investment but then you will realize that it will also open up many other avenues uh of value as you build um that kind of capability in your organization just as Cox's Automotive did the second scenario we're going to be talking about the supply chain optimization which um involves increasing the visibility into the endtoend supply chain to increase the uh demand forecast accuracy improve Warehouse Logistics and mitigate risks
for your organization so what are the challenges some supply chain intensive companies face well the first is that data is all over the place in terms of volume variety and velocity uh there's just so many different data sources with so many formats and coming in at different speeds that it's hard to capture let alone unify integrate and produce insights from it the second is improving accuracy at scale on a subfunction level you could produce a uh robust um ml algorithm that will help you with
your demand forecasting which is a key um which is a key part of supply chain optimization and planning the problem is how can you actually scale that to the whole organization so that you have that same view um across all functions and across the end to-end supply chain the third is that uh you don't have a lot of time to make decisions demand planners generally need to align with a lot of internal stakeholders across marketing finance and the supply chain itself and still uh make their releas
e dates which require a lot of lead time for many many reasons um so they don't have a lot of time to make these decisions and across the supply chain itself there is lack of visibility even if you could capture each data point by itself having them uh correlated with each other uh is a challenge the second is having too much noise um there's just so much data uh coming from different places uh you kind of have to find a needle in a hay stack and actually uh identify the important parts of the d
ata that are useful to you and and um just park the rest and the third is coordinating actions having everybody in the organization uh working in the same context uh in a way that is efficient that is agile and enables quick decision- making while also capturing the same view of what's happening in the supply chain is also a challenge so looking inwardly uh at Amazon as a whole Beyond just AWS uh um we've been building different Innovations across the last 20 plus years in our own supply chain u
m so that we can uh do all of these address all of the challenges that we just mentioned and at the heart of them of course is forecasting forecasting is super important in supply chain planning because under forecasting errors result in missed opportunity and over forecasting result in wasted resources and when we started this journey in the '90s we were using basic statistics iCal methods that um were working very well for basic data sets and capturing simple Trends and patterns but they ran i
nto Supply they ran into limitations as Supply chains became more complex and product assortment became more varied so fast forward on uh into the past 2015 and we and since then we've been using uh deep learning and sophisticated machine learning algorithms to match the complexity um of our our current supply chain and have really granular uh forecasting capabilities across Regional and National levels across different seasons um being able to correlate relationships between products look at sl
ow moving versus fast moving products look at the effect of pricing and promotions and the effect of external events uh on our demand planning process and we've taken that experience and as AWS we've helped uh customers realize similar values from those kinds of capabili abilities where we've invested those 20 plus years for example more retail one of the largest Hypermarket and Supermarket chains in India is using um our forecasting capabilities to reduce uh food waste and through more accurate
accurate machine learn based forecasting they were able to reduce waste from fresh produce by 30% and improve instock rates from 80% to 90% as well as sense Demand with external factors plan for new products meet demand from pricing promotions um uh optimize their Staffing decisions and um of course the actual reduction in waste that we mentioned earlier The Innovation from from an Amazon as a whole perspective didn't stop at just the forecasting capabilities um as you may know uh maybe you've
seen videos before online of our fulfillment centers and the processes that we have here have there we've invested heavily in robotics and automation uh on into the end to-end process and uh We've realized a lot lot of uh business value from that kind of automation such as reduction in fulfillment costs Improvement in inventory throughput uh increased labor productivity improvements in space utilization Freight spend and ab inventory accuracy as well as the movements and the efforts of the the o
ur uh Frontline workers in the Fulfillment centers themselves taking all that experience together um we're proud to have announced in rein at reinvent in November 2022 the preview of our uh AWS supply chain solution and our AWS supply chain solution allows you to do a couple of things number one easily connect data across systems addressing that first challenge uh harmonizing the data into a unified view viewing uh insights and risk alerts powered by Machine learning algorithms accelerating uh m
itigation for those kinds of risks with recommendations provided by the platform and of course like any great Cloud offering uh we aspire to provide you with um a flexible elastic pricing model uh with pay as you go uh pricing to take an uh a screenshot of what uh supply chain looks like here you can get uh 30,000 view of your inventory map across uh your different uh warehouses or fulfillment centers and you can drill down uh to inventory for a specific location where you can see your onand you
r safety stock and your uh stock in Transit you can take that a step further to start flagging uh any kinds of uh watch lists for Overstock and under stock risks and the supply chain solution itself will surface ml insights and potential risks like stockouts or overstock and it will also help you predict accurate lead Times by identifying which places uh where real world lead times are differing from the assumptions that you've built into your planning models and then based on that it can also r
ecommend different actions uh that that you can take and it will rank those uh actions based uh on different scores and you can choose what uh what option uh is more suitable for you in addition to the score you'll also see an estimate of uh improvements or effects or impacts on the kpis that are being flagged as a risk or a candidate for mitigation or action like uh inventory risk or uh CO2 emissions now we also mentioned that it's difficult to get uh lots of different organizations collaborati
ng in the same context and this is why we've built this kind of uh collaboration module with inbuilt chat and um identity where you can have different people from uh different functions uh looking at the same view looking at the same recommendations understanding the same risks to your supply chain so that they can take action faster now we mentioned the the importance uh of forecasting and we've taken the insights we've realized from our journey in building forecasting capabilities across the A
mazon supply chain into the AWS supply chain solution so you can use it in your demand planning and this will allow you to generate highly accurate demand forecasts at lower costs and um it uses machine learning to remove the manual effort and guesswork from demand Planning by delivering forecasts that are more accurate than traditional methods methods and you can uh create visualize and collaborate with different users to generate a demand plan uh you can also uh look at different metrics that
are related to this demand plan you can override the forecast uh even though it is ml-based you can explain the forecast as well so that um in your collaboration process with different stakeholders you can finally reach a consensus on uh a demand plan so um utilizing AWS supply chain you can quickly gain visibility on your supply chain end to end you can make more informed decisions resolve issues quickly and mitigate risks and uh lower costs the two key use cases we just talked about customer 3
60 and supply chain optimization require um changes to be made in the organization in terms of um platform technology and people and processes as well but the benefits that you can realize from such a change are not limited to those use cases alone and in fact adopting a datadriven decision making culture across your organization by itself across any project yields many benefits and in fact highly datadriven organizations are three times more likely to report significant improvements in decision
- making compared to those who rely Less on data and yet the current state is that the majority of Executives rely more on experience and advice than data to make business defining choices and so again we Face the challenge of what can we do to close the data value Gap well there are a couple of steps you can take to make your organization data driven and have your Executives uh trust that decision uh trust your decision-making engines and build that into uh the rest of the organization's decisi
on-making processes and the first starts with fostering a data driven culture which means that culturally you need to treat data as a strategic asset and then build capabilities to put that asset to use not just for big decisions but also for everyday action on the front line taking it from just the top and all the way down to the bottom and the people who are in operations marketing Finance sales where it becomes part of the lifeblood of the organization and um to do that you need to also democ
ratize access to that data and that means moving towards more modern data architectural constructs such as data mesh and data Fabrics to offer flexibility enable seamless and efficient access to data and govern that data usage across different business units another way to democratize access to insights is to stop requiring people to stop uh to uh stop what they're doing and access a different tool for insights instead what you can do is build analytics capabilities directly into their workflows
with context so that they can easily apply the insights to their daily activities taking that a step further building self-service analytics capabilities instead of just having a central function always service all requests is another important way of democratizing access to data and um and an an a vision for the future that we're building towards now is having zero ETL uh technology offerings as well as lots of low code no code uh technology offerings as well that allow the data analyst the ci
tizen data scientist and all any type of business user who is interfacing with data to use an intuitive easy to use no code or low code interface to analyze that data as as they are the do main experts and they understand the context and this reduces the Gap and the friction between surfacing insights from the data and actually putting it to use in uh the domain that they're working in similarly providing business intelligence tools for all users where they can uh uh surface the insights right a
way Beyond just accessing the raw data and having especially for repeatable processes that um single view or single pan of glass for each of their respective businesses so they can understand what's happening in the past and use it to make decisions for the future and use ml powered insites to predict uh the future and prescribe Solutions as well finally all of this needs to be underpinned uh with uh robust data governance and in fact some organizations are slow to open their data to their emplo
yees because they're concerned about this now what we're saying here is that this is a challenge that needs to be addressed headon and if it is then you can build that right granularity giving the right uh users the right access to the right data at the right time um and that will unlock all of these other capabilities that we just mentioned putting this all together um we'd like to talk about Coca-Cola andena which is Coca-Cola based in Chile they um decided that they want to become a datadrive
n organization and they've gone through that Journey of collecting all the relevant information they have on their company customers Logistics uh coverage assets to actually embark on that journey and they not only looked at the tools and capabilities of AWS um to uh to accelerate that Journey they've also looked to our uh Professional Services team to help them with training upskilling um and building out the capability in the organization and in fact uh for more than 300 hours of training prov
ided by AWS they were able to acquire these necessary capabilities to become a data driven uh company and they've increased uh productivity and eff and efficiency in decision- making across all different areas of the business so the outcome of this is that they were able to ingest more than 95% uh of the data that they uh needed from different areas of the business and they increased the productivity of the analytics team by around 80% now up until now we've talked about cases where you're inges
ting the data and you're analyzing it and then producing value from it but this is not the only way to realize business value from data and in fact you can actually build in what we call intelligent applications by embedding the analytics and the AI and the Machine learning insights into uh your applications themselves as they run uh into their day-to-day interactions with your businesses and your processes so in a WS we've looked across um a lot of common use cases where machine learning is use
d and we've used best-in-class algorithms and our uh internal uh data science and research and product engineering teams to build uh robust algorithms to address each of the use cases that you see here on the slide and we've packaged them up into reusable uh pre-built services that you can embed into your applications right away without necessarily having to have um data experience unless you're using pure machine learning services but for example for things like content moderation and fraud pre
vention and forecasting all of that is ready for you to use provide the right inputs and outputs from it and you can uh provide that intelligence uh straight away into your application and United Airlines has done just that where they um have created a loyalty program where they aim to maximize engagements with their with their customers and they use this to provide bpos spoke personalized offers uh to the customers using the Loyalty program and they succeeded and saw uh much more engagement ove
rall in the app especially among Millennials uh who they were targeting and they saw an uptic uh based on uh embedding intelligence into their applications so we talked about different use cases that will hopefully help you improve your top line but to maximize Roi we need to talk about the bottom line as well and that means keeping your costs lean so we're going to be talking about how AWS can help you optimize your costs one way we help you do that is to optimize the cost of uh scaling informa
tion infrastructure and we do that in a couple of ways number one is by offering elastic scalability capabilities through our services what that means is you can scale up or down or out or in depending on how much additional capacity you need to add to your services to match an uptick or a spike in demand and then lower it once that demand is lowered again so that you're not over provision or underprovision visioned another way the second thing that we um help with in that regards from the infor
mation infrastructure perspective is providing managed services and service level agreements so that we manage a part of your technology stack that has to do with managing the infrastructure and you can focus on the part of the technology stack that adds value to your business the third way that we do this is to optimize for Price performance we build in optimizations for Price performance when we are designing and tweaking and modifying our services so for example if we talk about our data ware
housing solution we offer 3x Better Price performance than other cloud data warehouses another way we help with the the optimization of cost for the information infrastructure scalability specifically is providing different mixes of on demand and reserved instances so if you um are not yet aware or comfortable with uh the predict ility of demand for a certain service you can use it in an on demand pricing model and once you have a steady state workload that has predictable demand you can then uh
further optimize your cost by reserving instances and getting further discounts for that the second way we can support is by optimizing data integration costs so just like with data warehousing our data integration services are can be up to 1.6 times faster for example uh than other um third party managed services with up to one tenth of the cost and again this is something that we continuously improve in addition we offer a variety of data integration services like managed open source managed
proprietary visual interfaces script-based interfaces notebook interfaces fully code based or low code or no code based so you can optimize your engineering effort you can optimize your licensing cost and you can optimize uh your mix of Technology choices that will will produce the best um um price mix for your organization and your use cases we also um have a vision of having a zero ETL future as mentioned before and having a zero ETL future and adding Integrations between different analytics S
ervices means you can save on the compute cost of building ETL and and the engineering cost of having to manage and build and fix and maintain uh several L pipelines the third way uh we can support you on optimizing your cost is to upskill staff to improve productivity and that involves training and certification it Evol involves uh immersion days workshops and Hands-On Labs that we can support your team on we can help with prototyping experimentation and design teams that can co-create assets w
ith your engineers and data scientists and data and uh data analysts we have our Professional Services team that can help with end to end strategy or implementation of um uh complex projects or programs and we provide significant Investments uh for funding for migration for modernization and for transformation projects as well so to recap there is um uh a need for organizations to build data driven capabilities however there is a gap between having the data and realizing the value from it closin
g that Gap requires a assessing uh the highest value use cases for your business thank you for um staying with us for the whole session uh hope you enjoyed it hope you realized some value from it so please complete the session survey and uh thanks again for listening these are my contacts happy to discuss with you specific details about your data Journey thank you
Comments