Microsoft PowerPoint - CGAP_customer analytics_presentation_final

[Audio] Customer Analytics Data analytics for customer centric business approaches CUSTOMER CENTRICITY SERIES | DATA ANALYTICS MODULE.

[Audio] What industry are you in? A. Transactions B. Insurance POLL CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS B. Insurance C. Savings D. Credit E. I work across financial services F. I do not work in financial services.

[Audio] Would you say your organisation is A. Developmentally focussed POLL CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS A. Developmentally focussed B. Profit focussed C. Somewhere in between the two.

[Audio] It is useful to start by defining the 1. What is your top challenge? 2. What does this challenge look like from WRITE DOWN: CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS defining the business problem 2. What does this challenge look like from the customer's perspective? Feel free to post your challenge in the questions box.

[Audio] Based on your business problem, what additional customer-related questions do you have? Customer RETENTION Customer ACQUISITION Customer INTERACTION Customer EXPERIENCE, PROTECTION & EMPOWERMENT CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS • Who are our new customers? • Did they engage with a sales campaign? • What is our cost to acquire new customers? • How can we improve the customer acquisition process? • Do different customer groups/ segments behave differently? • What are the preferred transaction and servicing channels for customers? • How can we better serve different customer segments? • How satisfied are our customers with our current offering? • Are our sales practices in the best interests of our customers? • Are we selling the correct products to the right customers? • How actively do our customers use our products? • What are inactivity or dormancy rates? • Which customers/ customer groups are terminating? • Why are they terminating? • How can we reduce churn?.

[Audio] Now what? Find Draw Assess CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS Find data Draw insight Assess impact.

[Audio] 1 Find data CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS • Types of data • Data challenges • Data privacy.

[Audio] When thinking about data, consider the full spectrum, from thick qualitative data to big data in the form of administrative and transactional data Data may already exist (secondary data), or it may need to be BIG CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS may need to be collected (primary data) BIG DATA THICK DATA.

[Audio] Data on your customers may be collected at different points along the customer journey Data collected through PRODUCT USAGE Data collected at CUSTOMER SIGN-UP Data collected through ONGOING ENGAGEMENT INTERNAL DATA CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS KYC data On boarding data Transactions Balances Channels Termination Sales campaigns Education / awareness campaigns Loyalty programmes Customer satisfaction surveys Customer outcomes or impact assessments.

[Audio] There are numerous challenges with regards to accessing 1. The data is not easily accessible 2. We do not have a single customer view 3. Data quality is poor CHALLENGES: CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS to accessing and preparing internal data for analysis 3. Data quality is poor 4. We have too much data 5. We operate in a low touch environment and so we have limited data on our customers.

[Audio] The data is not easily accessible What is a data warehouse? - A centralised repository of an organizations data • Data is organised and stored in such a way as to support forecasting and decision Many organisations do not have a data warehouse. Data on product usage might only sit in a source system Source systems are designed to support typical day to-day operations. They only provide a current snapshot view and the data is regularly overwritten CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS • Data in a source system is not designed to support decision making • Drawing data from a source system can pose operational risks as to support forecasting and decision making processes • Historical data is stored • Data imported into a warehouse is generally read-only. This means that once loaded it rarely changes thus preserving an ever-growing history of information • Data stored in a data warehouse can be accessed easily snapshot view and the data is regularly overwritten If developing a data warehouse is not viable, interim solutions include taking periodic snapshots of data held in your source systems and building a database over time.

[Audio] We do not have a single customer view If you're interested in your customers, you need to be able to analyse your data by A single customer view enables a holistic view by linking various data sources via a unique identifier that matches and brings together all data on a single customer Why is it important? CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS customer and not by transactions, accounts, policies or claims It allows your organization to understand how customers behave across multiple touch points KYC data On boarding data Transactions data Balance data Campaigns data Communications data.

[Audio] Nigeria example form the Nigeria Interbank Settlement System (NIBSS): The traditional view - volume and value of transactions Volume of transactions (Thousands) Value of transactions (Naira Billions) Number of transactions (Thousands) 83,915 50,000 60,000 70,000 80,000 90,000 8,395 5,000 6,000 7,000 8,000 9,000 Volume of transactions (N Billions) CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS Number of transactions (Thousands) 20,061 7,947 31,927 0 10,000 20,000 30,000 40,000 50,000 Jan 2017 Mar 2017 May 2017 Jul 2017 Sep 2017 Nov 2017 Jan 2018 Mar 2018 May 2018 Jul 2018 Sep 2018 Nov 2018 4,088 91 272 0 1,000 2,000 3,000 4,000 Jan 2017 Mar 2017 May 2017 Jul 2017 Sep 2017 Nov 2017 Jan 2018 Mar 2018 May 2018 Jul 2018 Sep 2018 Nov 2018 Volume of transactions (N Billions) NIBSS Instant Payment POS NIBSS Instant Payment POS Source: NIBSS published statistics.

[Audio] The Bank Verification Number (BVN) in Nigeria is a unique customer identifier that can be used to analyse transactions data at a customer level NIBSS Instant Payment customers POS customers CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS Source: Research conducted with NIBSS and Insight2Impact see: https://i2ifacility.org/insights/publications/advancing-financial-inclusion-executive-summary-nigeria-pilot-study?entity=blog.

[Audio] Different datasets can be connected using the unique customer identifier to gain a holistic view across transaction platforms CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS Source: Research conducted with NIBSS and Insight2Impact see: https://i2ifacility.org/insights/publications/advancing-financial-inclusion-executive-summary-nigeria-pilot-study?entity=blog.

[Audio] Privacy is critical In the first 6 months of 2018 there were 945 data breaches with 4.5 billion data records being compromised CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS https://www.cbronline.com/news/global-data-breaches-2018 https://www.informationisbeautiful.net/visualizations/worlds-biggest-data-breaches-hacks/.

[Audio] • Ensure data is stored securely • Create a data standard that specifies responsibility, Privacy is critical In the first 6 months of 2018 there were 945 data breaches with 4.5 billion data records being compromised CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS • Create a data standard that specifies responsibility, accountability, policies, and procedures on data sharing and analytics • Mask unique identifiers using hashing algorithms https://www.cbronline.com/news/global-data-breaches-2018 https://www.informationisbeautiful.net/visualizations/worlds-biggest-data-breaches-hacks/.

[Audio] Data quality is poor Not standardised Not verified What can you expect from the data? • Transactional data that is recorded per transaction is generally highly reliable • Data provided directly by customers can be less reliable CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS Not verified Not updated Not complete What can you do? • Map out the data you collect from your customers at each customer touch point • Prioritise the data you need to answer your specific customer questions • Improve data collection for those key data points as far as possible.

[Audio] We have too much data Some organisations have so much data it is difficult to know where to start Go back to your customer challenge and the related customer questions Stay focused CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS Use sampling Sampling can help make un-manageable data manageable . A well drawn random sample will produce highly accurate results.

[Audio] We have too little data Organizations that operate in a low touch (or low cost) environment often have very limited data on their customers There are numerous options for these Collecting data from customers may not make sense in many organisations that are purposefully low-touch to keep costs down CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS  Partnerships  Internet of things  Loyalty programmes (behaviour modification)  Social media limited data on their customers There are numerous options for these organisations to gather data:.

[Audio] Example - Discovery CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS 21st South African Financial Services Conference BRETT TROMP| CFO| DISCOVERY HEALTH https://www.discovery.co.za/assets/discoverycoza/corporate/investor-relations/ubs-financial-services-conference-october-2018.pdf.

[Audio] Rate you internal data assets POLL A. High - My organisation has access to data on our customers that is of a good quality CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS A. High - My organisation has access to data on our customers that is of a good quality B. Medium - We have some accessible data on our customers C. Low - My organisation has limited data or poor quality data on our customers.

[Audio] Draw insights and assess impact 2 CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS • Analyse data • Distil insights • Share insights.

[Audio] Once you have the data, the next step is turning the data 1. Build the dream team 2. Start small 3. Experiment Drawing customer insights: CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS turning the data into information and information into insights 3. Experiment 4. Beware of shifting the focus off the customer.

[Audio] Build the dream team The team lead is the customer insights champion in your organisation. This person will help drive the process internally. She or he should be curious about customers, able to identify TEAM LEAD Has the technical knowledge and skills required to manage multiple data sources and extract the necessary data for analysis. Has the best understanding of which data is available, its quality, and how easy it is to access. Could be an IT person, database administrator, data analyst or developer DATA PERSON/ TEAM CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS about customers, able to identify interesting questions and creatively identify the best available data to answer them, and can manage various teams in different areas of the organization administrator, data analyst or developer While business leadership has a vision of what they want to know about customers, the analytics team assesses what is possible given available data, runs the actual analysis and shares findings with the team so that useful business insights can be extracted ANALYST/ ANALYTICS TEAM.

[Audio] Start small Descriptive analytics Diagnostic Optimization AI Simple customer metrics and analytics, such as aggregating and summarizing customer data, can have a strong business impact CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS Predictive analytics Machine learning  There is a much higher chance of success if you start with manageable tasks  Basic analysis also helps you explore existing data for data quality issues  Simple metrics and basic analysis can spark new, more complex questions.

[Audio] Experiment Act on customer insights generated through analytics by directly testing and experimenting solutions with customers • Once customer insights have been generated this should be carried through into the business strategy and/ or operations • Experiment with the customer value CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS • Experiment with the customer value proposition, servicing, product design and optimization • Once any change has been made, keep analysing new data generated to assess impact on customers • Often requires trying and failing, and continuously adjusting for improvement.

[Audio] It's easy to get sidetracked and shift your focus from customer-centric to product-centric analytics DATA Be careful that you do not shift focus from the customer Continuously assess impact on the customer CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS IMPACT DATA INSIGHT Be sure to stay focused on customers.

[Audio] What is possible in my organization and how can we 3 CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS organization and how can we get there?.

[Audio] What can you do to get started on the customer analytics journey? What does this look like from the customers DEFINE YOUR BUSINESS CHALLENGE FIND DATA What data does your organization hold? DRAW INSIGHT Who would make up your dream team? ASSESS IMPACT Has the customer's experience been CUSTOMER CENTRICITY SERIES | CUSTOMER ANALYTICS from the customers perspective? What customer related questions do you have based on this challenge? organization hold? Are there any useful alternative data sources you could access? your dream team? Can you start small with some simple metrics? How has this impacted the business? experience been improved? Read the guide: http://customersguide.cgap.org/learn-from-customers/analyze.