1. Big data has the potential to significantly increase operating margins and productivity for retailers.
2. Retailers are investing in big data to improve merchandising, marketing, e-commerce, supply chain operations, and store operations.
3. Getting started with big data requires determining current maturity, identifying high-value use cases, assessing data and analytics capabilities, establishing data management processes, and anticipating business changes.
Hello everyone..I appreciate everyone’s attendance and sincerely hope that you gather some very valuable information and insights from our presentation today. This is a very exciting topic in regards to Big Data within Retail.We are going to spend time in this webinar on why big data is so bigthe reasons why big data is critical for retailgetting started with big data. 5 steps to assure that you are on the right trackhow big data can transform retailimplementing big data within your organizationas well as a few use casesLets get started.
So why is big data so big? What happened?In the last 2 years, statistics from various publications state that 90% of the world’s data has been created since 2010. With the emergence of smart phones, social media, and both user and machine generated data, data is growing exponential in size.Wireless devices are increasing data volumes as they outpace traditional devices. At this rate, we will begin to see close to a 35% year over year increase for the next 3 years. My prediction as new business models emerge, we will begin to see 55% year over year increases. Definitely setting a standard in zettabytes rather than petabyte storage.Digital content is also increasing data volumes immensely due to product models capturing more precise indicators. Products by YouTube, Netflix, Amazon, and others are changing their business models in the next few years to create more detailed revenue streams.
Big Data is critical in retail just to keep up with the masses.In most retail organizations, internal data is very challenging to comprehend in understanding your customer as well as demand.Publications state that 1/3 of retailers are in the dark regarding data that could be available to them. The Silo approach within organizations is the primary cause of the broken data pipeline.Primary reason as of why this is a hurdle are due to:*The lack of sharing data – major obstacle in measuring ROI*Misuse of available data in marketing communications – not able to personalize to your customer*Linking data at the customer level – thoroughly understanding use behavior*Infrequent data collection – only extracting what is needed within your traditional reporting ecosystem*Not enough customer data – not capturing the details of the customer (includes proper timings of viewed product, key indicators on why a user looks at one product versus another and so on)
Most retailers definitely understand the value of Big Data, but here is an interesting breakdown of actual improvement versus deployment expectations.As any retailer, having merchandise is quite important to sustain your customer base, but marketing is equally important. If you look at this slide closely, you will notice the ordering within improvement versus deployment. Overall, communication becomes the number one reason in actual deployment. The reason? To keep the customer coming back. And hopefully not in an annoying way.From my experience, all of these are important to keep the customer happy, committed, and loyal. The Big Data ecosystem allows you do so…
So how does an organization get started with Big Data? Customers today are extremely demanding and companies are becoming more savvy in user behavior analytics.. How should you approach this?You can get started with this 5 step plan1-How mature is your organization in it’s approach to big data? There are several types of POC’s to guide your ongoing investments. POC’s can consist of various types of projects to ease yourself into the big data ecosystem.2-Understanding your business functions. Big data can drive improvement and create detailed use case scenarios. Key areas to start with are pricing, segmentation, and marketing effectiveness.3-Size up your data management and analytics capabilities. Identifying your gaps and derive a game plan for project completion.4-making sure your data plan utilizes data management, internal policy and process governance, as well data collection for using and sharing your data.5-Anticipate process change and helping teams adjust to incorporating big data analytics into precise decision making.The 4 v’s volume, variety, velocity, and veracity will incorporate a proper thought process in tackling these 5 steps.
Retail organizations today have tried to take on the one-to-one market approach. This entails delivering to the right customer at the right time in the most relevant way in communication. But as you know, you are constrained by customer segmentation.As you can see in the flow in todays marketing approach, it allows you to follow up with a customer based on a campaign that you have launched. Big Data technologies allow you to dive deeper in understanding the customer user behavior and providing individual communication. A much more detailed approach in targeting, measurement, and performance.
Within your organization, there is not a uniformed approach in answering unique questions, but Big Data technologies allows you to minimize risk and increase likelihood of successful outcomes.Implementing Big Data should be approached utilizing these points:*Begin with Stakeholders – Who are the stakeholders and what is the criteria for success? These individuals are the knowledge workers and the decision makers.*Consider your Culture when it comes to Big Data – This involves a cultural shift which expects data-driven, fact based decision making. Gut feel conclusions are not supported.*Finding your data stewards – Individuals with a mix of technical and business skills. This can be a single person or members of a team that will produce successful results.*Set clear goals with Big Data – Don’t boil the ocean! Start off small and get an understanding of how you can utilize your Big Data ecosystem.*Create your plan – Link your company goals to the 4 V’s. How can you solve your questions that cannot be solved today.*Establish metrics – Limit these to high priority measures at the beginning. This allows your organization to move away from the silo approach to analytics business decisions.*Deploy the technology – Proper architecture and ecosystems that allow a uniformed data environment.*Making Big Data Little – Delivering little data in context to business use cases. Overall, showing the value of your big data ecosystem.
In this high level solution architecture, you can see how a Big Data ecosystem is a separate solution from your current reporting architecture. Real-time streaming analytics and your Big Data refinery can store structured data from your online serving system, but also includes unstructured data, social media, wireless data, etc.This allows a streamlined approach to analytics and reporting..Notice that the Big Data refinery of unstructured data and real-time streaming can be aggregated and utilized within your reporting ecosystem. This allows combined aggregated dashboard reporting for both structured and unstructured data.Another use for a solution like this is the storage and availability of data. Allowing for more flexibility in A/B Testing, User Behavior analysis, New Product testing and optimization, fraud detection, as well as an environment for your data wranglers.You can also add API’s to deliver data to your clients or create new revenue streams within other products. The innovation efforts are endless.
Pactera can provide a step-by-step Big Data Solution throughout the Lifecycle:We Offer scope of services including a 4 hour workshops, 2-4 week proof of concepts, and project implementations. Projects can include benchmark and monitoring, Integrations and migrations, Implementation and solution architecture, project management, analytics, and reporting within your Big Data solution.
Pactera senior consultants will facilitate a focused workshop with IT and business leadership to deliver a customized “Big Data: What to do Next” guide. Our consultants will outline a step-by-step process to building a Big Data solution, including requirements and specific benefits.Pactera Executive Workshop is based on Strategy, Planning, and Expectations within your Big Data Solution.Workshops include:*Big Data Strategy on what your tomorrow will look like*Establish Market Dominance*Project takeaways including POC’s, analytics, solution architecture, reporting, and monitoring*Roadblocks in implementation*Defining your ROI*Getting the right ROI from your solutionThis workshop is geared towards executives and management explaining the long term value of a Big Data ecosystem..
The Pactera technical Workshop is based on: end to end managementSystem tuning/auto-tuning and configuration managementDealing with both structured and unstructured dataMonitoring, diagnosis, and automated behavior detectionSolution architectureProcessor, memory, and system architectures for data analysisBenchmarks, metrics, and workload characterization for big dataAvailability, fault tolerance and recovery issuesData management and analytics for vast amounts of unstructured dataThe technical workshop is geared towards technical architects, developers, operations, and end users.
Now we will move into a few use cases for Big Data in RetailFirst we have the use case associated to customer insight and behavior
Retail analytics usually follow this metric flow depending on your retail model. All of these are measured in some form or fashion either within your current reporting ecosystem or from a 3rd party solution.But is this really answering all of the necessary questions? Maybe at a high level to detailed aggregate level it is..Does this answer all your business questions? No, not the what if’s..
Here are some questions that can be solved outside of your structured ecosystem. The proof is within the data.In this use case, the Big Data Refinery is helping this client analyze some extremely important viewpoints to understand their customer:Understanding their behavior and preferences.>>testing rapidly and building user behavior models of particular customers>>utilizing structured and unstructured data to complete the modelsStructure and Analyze Data on their Consumers>>Social graphs and traverse usage patterns of the consumerFeature extractions to find root causes>>Defining new attributes and model statistical significance>>Combinations and sequences of consumer attributes and actionsAll important to understand the lifespan of the customer and their behavior.
Another scenario within this use case is customer loyaltyComparison shopping is making retail purchases extremely hyper intensive>>this means it is important to understand your customer at a more detailed levelOffering discount programs, detailed email correspondence, and developing brand loyalty with your customers.Customer lifecycle is more than just purchases>>Retaining the customer is quite challenging. Taking to the next detailed level involving browsing of products to capture customer attention as well as bridging the gap between their purchasesReaching into online channelsAnother detailed level that can be accomplished is making their online engagement as personalized as walking into a store. This will show customer awareness and a loyal following
This process flow shows you how to handle the previous scenarios within their Big Data refinery. In this case, the refinery is Hadoop. The intake data was a combination of structured and unstructured data on their customers. Thus, Allowing them to analyze additional metrics and combining the sources within their current reporting ecosystemThe end result is within their current reporting repository.
Next is a use case on brand and sentiment analysis. In this case the client is interested in Brand Recognition
It is important to understand what is being said about the brand, Social media generates a flood of data. Especially within a larger brandCapturing and processing direct feedback within this client>>better engagement means accurate sentiment analysis and allowing integration within their customer service systemsHandling diverse data types and associated processing>>Data sources can and will change and the semantics are always continuously evolving>>Their refinery also allowed vast improvements within current algorithms as well as testing of new scenarios.
This flow shows you how they handled brand recognition within this large retail clientIn this case, the refinery is Hadoop including the real-time in-memory solution Hbase. Your intake data was a combination of structured and unstructured data on your customer. The end result is providing trend analysis and utilizing a search GUI and reporting applications to your partners. Yet, another revenue stream.
Thank you for attending Pactera’s webinar on Big Data within Retail.Are there any questions?**cricketsHere are a few questions that were asked during our presentation:Q:When you refer to the Big Data Refinery, is that really just Hadoop?A: no, it can be any distributed platform that your organization feels comfortable with that is scalable. Hadoop is only 1 example. Our clients use MongDB for document oriented real time architecture as well as other no-sql distributed platforms such as Hbase, Storm Project, Cassandra, Redis, and the Spark Project.Q: Do we have to provide you with in-house hardware to conduct the POC and the project lifecycle?A: Absolutely not, we offer cloud services as well as a right shore model. We can work with you to fit your needs and take your organization to the next detailed level.Well, thanks again, both Challen and I on behalf of Pactera appreciate your time. If you do have any follow up questions, feel free to contact us!