January 19, 2020
- Posted by: admin
My last article talked about the Challenges we faced as a newly-minted Business Customer because StarHub lacks a Single Customer View.
This is a classic example of an Enterprise that has a multitude of different applications that are all managing customer data, and no formal process to update these systems when new records are received. I don’t know their exact IT landscape, but it’s reasonably clear they must have:
1. A finance/billing system to raise invoices to clients and register payments.
- In our case it’s likely this system still had our old office address registered against our company name and UEN, as the invoice had been sent to the wrong address
- Given we had registered for these new contracts with our latest BizFile, there’s no excuse for not updating this information in their Billing systems
- Given we could not pay the Enterprise bill in a retail store it is also possible they have more than one billing system in operation, meaning the record they have for me as their retail customer is not matched against the record they have for me as an Enterprise customer under Cognopia.
2. A system to manage Retail customers
- Sayonnara had been a retail client when using her FIN, and this had not been updated since she received Permanent Residency in Singapore
- Given we registered this new phone contract using her NRIC number, there is clearly no mechanism to match existing retail contact records with new info from Enterprise contacts, or to update the retail system, or to have a single view of Sayonnnara as a customer of both their Retail and Enterprise business lines
3. A system to display account details to the customer
- In our case this is their StarHub Business Manager application.
- The Account number shown in this application does not correspond with the number required to pay the bills. At the least this means there’s limited integration between systems to invoice/receive payment from customers and those that display contract details to those customers.
4. A system to accept customer support calls and route them to the right team
- In reality this was an automated line telling me I’d been cut off, so it’s not clear if there’s any system here. What was clear is the service did not register my mobile phone number and therefore cut me off without resolving my issue
How could StarHub do better?
Rather obviously given the opening paragraphs, my suggestion to StarHub would be to run an Enterprise Wide Single Customer View programme. Whether I’ve correctly guessed their systems landscape or not, it’s abundantly obvious that the systems they have that manage customer data are not integrated to one another and are not leveraging the same data.
The solution to this problem will cross business departments – Sales, Finance, Customer support. It will require Business Process Changes (so that there is some process to propagate new information throughout the organisation). And it can also involve IT systems changes. How might they do this?
The first step would be to identify and determine the data structure for a “Golden Customer Record”
- Which data attributes do they need to manage and maintain in order to serve their customers?
- What business processes exist along the customer journey and which data elements do they need?
- What format do we expect the data to be in? Defining a common format or list of acceptable values for each attribute will help. These Data Quality rules can be used to drive KPIs across the business and will be critical in demonstrating a “quick win” for the SCV project to Management stakeholders
- It’s really important at this stage to focus on core data that will deliver immediate impact. Your project can easily be de-railed here by scope-creep, attempting to bring through every data attribute from every system rather than those core components that deliver great customer experience
Start with Profiling
This can often be a daunting task, as different IT systems have different data structures and processes. At Cognopia we recommend leveraging a Data Profiling and Discovery solution – profile all of the customer data tables that exist in any IT system that manages customer data. Data Profiling can help us to discover value frequencies, formats and patterns that lead us to believe that the majority of our data reflects some basic standard. The benefits of profiling your data:
1. You get a comprehensive picture of the current state of customer data within your organisation – including checks on completeness, validity, uniqueness, timeliness and accuracy
2. Exposing the profiling results to your business users is an easier way to get their buy-in and engagement than showing them raw data in tables
3. You have a baseline set of data attributes from each source/target system from which to create your Golden Record attributes
Cleanse the Data
Modern Data Quality tools like Experian Aperture can seamlessly jump from Profiling to Data Quality remediation. If you’re embarking upon a Single Customer View project you should aim to cleanse the data before you go about matching and harmonising the records.
This step is where you’ll aim to standardise the existing customer data against the Data Quality rules you defined in the first step. You can also use this step to validate and verify key data elements such as Address, Email and Phone details – ensuring the data is both standardised AND valid will improve your marketing response rates in future.
You will want to capture the baseline level of Data Quality and record that against the Data Quality rules you defined at the outset in Step one – this shouldn’t be used to bash departments with poor data quality, rather it will act as proof that your remediation efforts are working and can quickly prove value to your management stakeholders once the project commences.
As you go through the cleansing process, you’ll be writing rules to remove incorrect, incomplete, duplicate and improperly formatted data from the sources. This process is an opportunity to perform Root Cause Analysis – trying to understand why your data degraded and then put policies, processes and training (Change Management) in place to prevent it from happening again.
Match and Merge
Finally we get to the stage where we harmonise the records and decide which data elements are going to be kept for the Golden Record. In this step we are going to identify relationships amongst the data records held in StarHub’s source systems in a bid to match those belonging to a single customer entity and merge the most valid data attributes together as the Single Customer View.
Simple matches are possible where there’s a unique client ID used across all systems, but it’s seldom that easy. More likely you’ll need to deploy complex fuzzy matching algorithms that will look for relationships using Jaro distance or “sounds like” phonetic matches. You can leverage customer address details, names, dates of birth, phone/email records etc to build up a match score and cluster together those records that score highly as duplicates.
- This step is often an iterative one, and you can benefit greatly from easy-to-use tools here as the business users will have a better understanding of the records that match – so select a tool that a business user can use!
Once you’ve got the cluster of similar records together, the final step is selecting the Golden Record attributes that are most correct to be promoted to the Single Customer View database (e.g. most recent, most populated, highest, lowest etc).
Rinse and Repeat!
There’s no use in treating this as a single exercise – Data Quality is an ongoing battle where we constantly need to measure, monitor, repair and prevent degradation of our data assets. Root Cause analysis is essential, and strong Change Management processes can be put in place once you’ve identified the reasons why your data has degraded. You also need to ensure the systems that need access to a Golden Customer Record are able to get it, and that this data is pushed to your target systems in a timely manner (in our case the billing system needed to be updated with the new address prior to the next billing cycle).
Additional processes need to be baked in to ensure the Golden Record is used by all systems – ensuring my invoice is sent to the right address, my phone number is associated with the right account, and the account number I have been issued in my phone app represents the same account number the finance team expect when I try to make them a payment.
This is not an uncommon issue to find across Enterprises of StarHub’s size and reach. Usually the systems selected and implemented by the sales and marketing teams will be customised and configured to make their life easy. The systems used in Finance will depend on good quality data captured upstream but may have different formats/constraints when holding customer data. Data Siloes exist and the business experiences friction as a result.
In many cases an organisation simply has not had the time to analyse the quality of their source data and the impact this has on the bottom line. The good news for StarHub is that there are great tools available to deliver this kind of project. With the right internal drive and executive sponsorship the organisation could easily address these issues, reduce customer pain, increase customer retention, and provide a rich repository of customer information for marketing to analyse and act upon to increase revenues for the business moving forward. The ROI on a Single Customer View project is often too big to miss.
How can Cognopia Help?
We work with industry leading experts and software solutions that are designed to make this process as easy, fast, and fool proof as possible. We can help craft the business case for a Single Customer View project to gain executive buy-in, as well as define the strategy, tactics, rules and tools you need to successfully implement a Single Customer View.
Our focus is delivering practical, pragmatic projects with tightly defined scope to unlock quick ROI for our customer base. If your business suffers from similar internal friction caused by poor data management practices, get in touch today.