Customer personalization is the way forward in this competitive business landscape. With giants like Amazon, Netflix, Spotify, banking on hyper-personalization, businesses are in a race to create personalized experiences to remain profitable. But there’s one problem – most businesses don’t have access to a consolidated customer view.

57% of companies in a research conducted by Marketing Week reported they have not yet built a 360 customer view.

76% of companies say that a single customer view is extremely important to their organization.

A customer 360 view, also referred to as a unified or a single customer view is a consolidated version of all your customer data scattered throughout the apps and systems in your organization. For most organizations, gathering data from disparate sources while ensuring its quality is a significant challenge that prevents a successful personalization initiative from taking place.

The very foundation of hyper-personalization is the consolidation of customer data – to deliver personalized services, you have to gather data from devices, apps, third-party vendors, systems and sources. Gathering data though is only part of a bigger problem; companies also have to deal with dirty and duplicated data, responsible for many failed digital transformation initiatives.

So if you or your team are planning to test a personalization initiative, this article can serve as a gentle reminder to ensure the accuracy and integrity of your customer data first before you use it for its intended purpose.

Understanding the Challenges with Customer Data

By aggregating user records from multiple sources, you create a centralized record that gives a complete view of the interaction customers have had with your company and the actions they performed at various touch points – on their mobiles, apps, on your website, or even in your brick-and-mortar store.

As simple as this sounds in theory, it’s a difficult process, partly because of how organizations store data and partly because companies are still caught in a web of mismanaged data processes.

It’s not uncommon for a company to use multiple asset management systems to store customer data. For instance, an enterprise with multiple departments may use Salesforce, HubSpot and Fusion to meet their respective goals . Some may even have third-party vendors who use their own ERPs or data systems to store customer information. Come the time for analysis or data review, decision-makers realize they don’t have access to complete information. Teams frantically import/exchange data to meet critical deadlines, while risking data loss and diminished data quality.

Take for instance, the case of an insurance firm launching a new health insurance plan for college-going kids of existing customers. To create this plan, the firm must have access to essential customer data such as family status, income level, health status and other relevant details. Chances are this data is already stored in multiple systems of the firm but because the firm never really took a serious in data preparation, found it extremely difficult to execute their plan within a specific deadline.

Because the firm does not have a data quality management framework in place, the data suffer from:

  • Duplication: Duplicates happen due to multiple reasons. A user could use a different email ID and sign up 3 times. A data entry operator may accidentally enter the same information twice. A data migration process may go awry. Duplication as shown in the image below may ruin data integrity as accuracy and uniqueness are compromised. The result? Skewed analytics and unreliable business intelligence.
    Instances of Duplicate Data – Data Ladder
  • Disparate Sources: When companies use multiple apps and systems to store, collect or process data, it results in disparate data – making it one of the leading data quality challenges that companies have a hard time fixing. The disparity in data leads to disjointed views, preventing companies from gaining access to a reliable truth.
  • Messy Structures: Data that have format issues, typos, misspelled names, incomplete or invalid structures can destroy the integrity and accuracy of information. If the data is manually keyed in and there are no defined standards or controls in place (such as using a drop-down to select a field), dirty data will be a significant challenge.
Poor Customer Data Quality – Image Source Data Ladder

Millions of dollars are wasted in return mails, lawsuits, expensive mistakes, inaccurate insights, and analytics – all because companies do not have a data quality framework in place to resolve bad data problems.

For this insurance firm, the first step is to make sense of its data. Once it has accurate, complete, valid information, it can then proceed to merge records to create the customer 360 view.

How do they achieve this?

Here are two approaches.

Two Approaches to Improving Data Quality

The firm in question can use two approaches to fix their data quality.

  1. The Manual Approach: They can hire a team of data experts to create an in-house solution to sort, merge, purge, and clean their data. This approach could take years. Hiring, training, testing, implementing solutions could take months of effort, with no guarantee that the data will be at least 95% accurate and complete. Manual methods work well if you have limited, structured, or semi-structured records or if there is a known accurate value (a unique ID, or serial numbers). With larger and complicated data sets, a manual method increases the workload of data analysts, enforcing them to focus on data cleansing rather than on data analysis.
  2. The Digital Method: There is no dearth of digital solution to resolve data quality issues. Several types of software can be used to manipulate data. There are native-solutions that are engineered to work with a specific data warehousing application. There SQL-based solutions that will require custom programming and a steep learning curve to work.And then there is self-service data quality management software that is designed to help employees access, clean, and transform data stored in multiple data sources. These solutions differ from generic solutions because they are designed for the business user and turns complex data quality processes into easy to perform tasks. These solutions use a combination of algorithms to match complex data, consolidate disparate data definitions, clean, parse, standardize, and homogenize data in one platform.
The data profiling and cleansing process

Whatever method you choose depends on your budget, data complexity, availability of resources, and the time you allocate to the project. It’s worth mentioning though that automated data quality solutions will help you cut down on significant operational costs while improving business performance. Your employees won’t have to spend hours manually fixing poor data lists. Your business won’t have to operate in the dark. You will truly be data-driven when you take data quality as the priority to a customer personalization initiative.

Using Data to Drive Customer Personalization Initiatives

By 2020, Gartner predicts that more than 40% of all data analytics projects will relate to an aspect of customer experience. (Gartner)

This is why, data, specifically high-quality, reconciled data is so important to drive customer personalization initiatives.

To personalize an experience, you must find out the hidden relationships between people, products, their environment, their behaviors, their experiences, their expectations.

Citing the insurance firm example above.

It took them 2 months to collect the necessary data from its vendors and partners. Next up, they had to clean, dedupe and normalize that. That process took them just three weeks when they used a self-service tool saving them considerable time and manpower. Finally, after 3 months of data cleansing, data match, merge and purge processes, the company was able to get a consolidated record of its customers. Alleluia.

The result?

They were able to segment their audience, create the perfect insurance plan for parents with college-going kids, ran effective email marketing campaigns, and were able to increase their ROI by 2X. Even better was the positive response from their customers.

So How Do You Get Started?

If you’re eager on running a personalization initiative, it’s recommended to start small. You don’t have to overhaul your company’s data. You just need to identify the type of data you need for the type of personalization service you want to offer. Some tips to help:

  1. Start Small: Create a process that you can manage. Extract 500 customer records, assess them for quality issues and note how much time it would take you to fix those problems.
  2. Decide on the Approach: Assessing your data for the three Ds will help you decide on the approach you want to use to fix it. If the data is complex with no confirmed accurate value and has 15 – 25% duplicates (that’s 125 bad records out of 500), you will need to use an automated solution instead of manually fixing it.
  3. Invest in the Right Tools: You cannot do this using Excel or Google Spreadsheets. If you’re pulling in data from multiple sources (CRMs, transaction data, behavioral data), you’ll need to invest in a tool that allows you to consolidate, dedupe data and clean up dirty data.
  4. Fix Your Data: I can’t help reiterating this. Bad data will be a huge bottleneck in EVERY aspect. Regardless of whether you want 360 customer view or not, you must make data quality a priority. A countless number of businesses have failed (with millions of dollars down the drain) because of bad data. So whether it’s for something as normal as a promotional campaign or for something as critical as annual reporting, you need data you can trust.
  5. Create a Master Record & Automate Cleaning: Once you’ve cleaned and de-duped your data, create the master record (also called the Golden Record). You can then use the tool of choice to automate the data cleansing. Further down the line, you can also augment this record with additional firmographic, demographic, or behavioral data, therefore, enriching your customer data.

Getting started may seem like a mammoth effort, especially if your organization’s data is not sorted, but with the right plan and resources, you can turn ordinary lists and records into a powerful intellectual asset – capable of driving your ROI, enhancing customer experience, and increasing business performance.

To Conclude:

I’ll keep it short. Want your customers to choose you over competitors? Give them the experience they need. But before you attempt to woo them, fix your data, give meaning to vague terms like ‘data-driven’ or, ‘data-centric,’ and be ahead of the curve.