Omni-Channel Customer-Centric Strategies in a Modern Architecture

Traditionally, the communications area of a software project in a big company was perceived as an utility, a mere delivery system for simple communications. Until a few years ago, it was considered almost miraculous to be able to send communications in real time.

In the year 2019 nobody should settle for that. We are in the era of disruption, specialization and complexity. The microservice-based and event-driven architectures are another mechanism to deal with that complexity by designing simple and specialized components. But how to achieve a completely customer-centric omni-channel strategy with an increasingly chopped technical substrate, very complex to orchestrate and prioritize?

In this talk we will tell how to design a modern communications architecture, based on microservices and driven by events. Some of the key messages will be:

- How to design a future-proof data repositories to have both performance and valuable insights of the customers
- The criticality of having a proper event strategy and clear data flows, in order to have the customer context and all the relevant indicators as fresh as possible without sacrificing the quality of the information
- The importance of integrating inbound and outbound to achieve the desired omni-channel perspective
- Why is it also key to synchronize the leads generated in the campaigns area into your real time ecosystem, even for operational purposes
- How to distinguish communications that must be in real time with those that can maintain its batch nature
- How to integrate propensity and prediction models into the strategy
- How to deliver this targets or segments of customers as an input for the rest of the platform
- The importance of integrating the feedback and responses from the customers, to reinforce a relevant dialog always
- etc.

A Big Data platform could help you in some of these important tasks, but first of all the overall design needs to be correct and some key patterns need to be implemented.

For the sake of clarity, we will show you real examples of how we are doing it at ING as well as the key elements that will guide our steps towards a more frequent and relevant interaction with our customers.

In short, we will give you some tricks to make any contact between the client and the platform as relevant as possible.

Transcript

1. Omni-channel customer-centric strategies in a modern architecture Big Things Conference Madrid, November 2019
2. Commercial Banking Retail Banking Why we are here 2 Luis García Castro @luiyo Pablo Ruiz Subira @prsubi
4. Introduction 4 Omni-Channel? Customer-Centric? Strategies?
5. Goal 5 relevant dialogue x-sell upsell advice help empower smart Omni channel customer preferences at the right time
6. Common Terminology 6
7. Terminology – Outbound channel 7 Outbound channels: • Email • Push • SMS • Postal mail • Phone call
8. Terminology – Inbound channel 8 Inbound channels: • Banners (web / app) • Content to be used in assisted channels
9. Quiz – Assisted channels nature: outbound or inbound? 9 Both! The customer or the agent can call in The agent has to have real time information to assist the customer
10. Terminology – Communication 10 Metadata Template Name Category Validity dates Description Operative Commercial Legal Advisory Offer / business proposition Promotion Static business priority
11. Main Data Repositories 11
12. Outbound contact history 12 ………. Metadata 1 Template 1 data Metadata 2 Template 2 data Metadata n Template n data channel channel channel outbound
13. Assisted channels contact history 13 assisted channels Contact history: • Planned callbacks • Incoming callbacks • Agent notes
14. Inbound contact history 14 inbound web app banners Contact history: • Web banners • App banners • Geo (web + app) • Assisted channel banners
15. Clickstream contact History 15 clickstream public website private website DMP / DSP 3rd party data providers Contact history: • Navigation track • Session ID • Action(s) tracking
16. Customer interaction 16
17. Contact integration status + Customer response 17 Bla bla bla push channel app banner postal mail integration status direct customer response inferred customer response push notification banner letter inferred responses definitions
18. lead Lead statuses and omni-channel synchronization 18 business proposition communication 1 communication 2 communication 3 push email callback
19. Most meaningful response 19 shown remind me later interested NOT interested accepted
20. Preferences and consents 20 Legal communications Operative communications Commercial communications Advisory communications Selected commercial communications Selected advisory communications Selected operative communications orchestration Consents Preferences
21. clickstream inbound outbound assisted customer profiles system of records customer analytics Integrated 360 customer dialogue 21 business strategy customer expert contacts responses machine learning and automated decisions control group …… control panel
22. Lessons learned 22
23. Lesson 1 – Vendor locking is a [unavoidable] consequence 23 Strong vendor lock driven approach Soft vendor lock driven approach Fully integrated solutions are great, and usually simple They provide good time to market but your final architecture should limit the impacts if eventually you need to change your partners The most crucial part is the DATA Set up your own definitions for communication, contact, channel and the rest of key concepts before any RFP Define your architecture Choose your solution accordingly before while doing after
24. Lesson 2 – Manage your Reference Data 24 Forgetting about proper reference data definitions is a common mistake in immature organizations Data exploration and usage must always require a minimum knowledge of each data domain To really enable a data driven strategy the reference data is a must. Set up your business definitions of those well-known discrete set of values and code them to a language agnostic set of values Without governing Reference Data Reference Data governed before while doing after
25. Lesson 3 – Enable data consistency 25 There is an important risk of inconsistency if not all the data assets are synchronized Do not overcomplicate the data synchronization, use a specific tool if it applies The simpler the better Weak data consistency Strong data consistency before while doing after
26. Lesson 4 – Define loosely coupled domains 26 The application domains fulfil vertical customer needs, end to end. There are silos. Wired connections between domains exist Try to identify clear domains and avoid the overlap among them Create different domains for cross functional use cases Poor domain definitions and blurred bounded contexts Strong domain driven design before while doing after
27. Lesson 5 – Choose wisely your real-time battles 27 You should not go from 100% batch data processes to full real-time since probably not everything occurs that way Do not sacrifice the quality of the data Data flows must always be clear The 360 customer context must have all the relevant indicators as fresh as possible Everything must be real-time There is a right balance between batch and real-time before while doing after
28. Lesson 6 – Set up a data driven culture in your organization 28 You are aware of the data importance but in the end data is just something needed to reach the customer Data responsibilities are a cross functional domain A real democratization of data is in place, under strong and carefully assessed design principles Data is what I need Data is the cornerstone before while doing after
29. Main components overview 29
30. Required components overview 30 Tool - GUI Event Broker Relational databases No-SQL databases Building blocks - Microservices Real-time decision engine High volume SQL engine Streaming engine Data exploration area Template engine Document Management System Preferences and consents Channel gateway Master Reference Data Reporting data asset 360 customer profiles Content Management System Big data platform Model execution engine GDPR Orchestrator CRM
31. Case 1: Basic outbound (i.e.: for legal communications) 31 Tool - GUI Event Broker Relational databases No-SQL databases Building blocks - Microservices Real-time decision engine High volume SQL engine Streaming engine Data exploration area Template engine Document Management System Preferences and consents Channel gateway Master Reference Data Reporting data asset 360 customer profiles Content Management System Big data platform Model execution engine GDPR Orchestrator CRM
32. Case 2: Simple outbound + inbound (operative comms) 32 Tool - GUI Event Broker Relational databases No-SQL databases Building blocks - Microservices Real-time decision engine High volume SQL engine Streaming engine Data exploration area Template engine Document Management System Preferences and consents Channel gateway Master Reference Data Reporting data asset 360 customer profiles Content Management System Big data platform Model execution engine GDPR Orchestrator CRM
33. Case 3: Simple outbound + inbound (commercial + operative) 33 Tool - GUI Event Broker Relational databases No-SQL databases Building blocks - Microservices Real-time decision engine High volume SQL engine Streaming engine Data exploration area Template engine Document Management System Preferences and consents Channel gateway Master Reference Data Reporting data asset 360 customer profiles Content Management System Big data platform Model execution engine GDPR Orchestrator CRM
34. Case 3: …. Including real-time simple decisions 34 Tool - GUI Event Broker Relational databases No-SQL databases Building blocks - Microservices Real-time decision engine High volume SQL engine Streaming engine Data exploration area Template engine Document Management System Preferences and consents Channel gateway Master Reference Data Reporting data asset 360 customer profiles Content Management System Big data platform Model execution engine GDPR Orchestrator CRM
35. Case 4: Scalable architecture with omni-channel, customer centric capabilities 35 Tool - GUI Event Broker Relational databases No-SQL databases Building blocks - Microservices Real-time decision engine High volume SQL engine Streaming engine Data exploration area Template engine Document Management System Preferences and consents Channel gateway Master Reference Data Reporting data asset 360 customer profiles Content Management System Big data platform Model execution engine GDPR Orchestrator CRM
36. Thank you!! Questions? 36