15 years ago, British data scientist Clive Humby coined the popular maxim data is the new oil which has since been updated by Ted McCandless to data is the new soil. We should all agree on the importance on data to support decision making as there are plenty of examples out there (Moneyball anyone?) but what about the cost of having bad data? 7 years in the CRM and analytics industry pushed me to write an article about this topic, adding some tips/recommendations to change its vision: from business cost to business opportunity.
Data quality dimensions
Data quality is measured by several dimensions. Merging knowledge from analytics leader SAS and marketing guru Neil Patel helped me list the main ones:
- Existence. Does your organization have the data to support the objectives?
- Validity. Are the values of your data acceptable and up to date?
- Accuracy. Does the data accurately describe the properties of the object it is meant to model?
- Consistency. When the same piece of data is stored in different locations, do they have the same value?
- Completeness. Does your data set have missing values?
- Uniqueness. Is a record represented once within a data set?
Given the amount of data available today, if you don’t put in place the best practice for your data quality management you could find yourself with a problem that grows exponentially. The following infographic from Neil Patel summarises it quite well:
Working on dozens of data related projects taught me marketing teams have plenty of information available, but it wasn’t until I googled for some stats to include in this article that I realized this opportunity/problem had a completely unexpected dimension. A couple of years ago Salesforce, the n°1 CRM software company, forecasted the median number of data sources marketers use to 15 in 2019. 15! This means there is a huge opportunity to deeper understand your customers and therefore be able to provide them with what they do want; but there may also be a huge problem if no best practice processes in data quality management are in place.
Let’s move to a real-life scenario. The marketing department of a football club is being asked to drive more sales for a specific product (say a match kit) on the club’s official online store.
Instead of sending a bulk email to each contact in the database, the marketing team decides to send a personalized email to all the following audiences:
- Audience A. All opted-in contacts who visited the “match kit” page on the online store twice or more but have not bought yet
- Audience B. All the contacts who have made a purchase, other than the match kit, in physical stores in the last 12 months but who haven’t registered on the online store yet. This information is provided by store owners as Excel spreadsheets.
What are the key issues with data quality?
Data quality issue #1 (Existence). Information about visits to the “match kit” page on the online store is available on an aggregate level and therefore the marketing team can’t use audience A in the campaign
Data quality issue #2 (Consistency). Merging multiple Excel spreadsheets to build audience B results in the discovery of a consistency problem: first and last name for the same email address are different and therefore the marketing team should spend additional time to decide how to manage the communication to these contacts.
Data quality issue #3 (Validity). While all store owners shared the list of customers who made a purchase in the last 12 months, one of them also noted there is missing information as he didn’t collect it in the last 3 months due to internal processes issues. The result? There may be customers who actually bought the match kit recently but the marketing team could not include contacts from this data source in audience B.
Given the scenario presented the following are true:
- Marketing performances are affected by data quality issues
- Marketing team spend a lot of time (=cost!) looking for the “single version of the truth”
- As of now, there is no “single version of the truth” within the marketing department as there is no Single Customer View (if you don’t know what a SCV is, let’s have a look at our dedicated article here)
A single vision to rule them all: Benefits of introducing data quality management
There are three main benefits of introducing data quality management best practice into a sports organization:
- Improve effectiveness: getting the right message to the right person at the right time means more sales, more engagement and the ability to meet fan expectations (if you are still not convinced of the value of personalization, you can find more information from McKinsey & Company, a trusted advisor and counsellor to many of the world’s most influential businesses and institutions, here)
- Reduce costs: people in your organization do not need to spend hours looking at dozens of Excel spreadsheets to create the right segment to send a personalized promotion
- Improve collaboration: sales and customer service teams do not need to phone marketing team as they can access independently the most accurate information about sponsors/fans
Start organising your data today
Every organization has its own pathway to success, and this is not different when working on best practice processes in data quality management. Nevertheless, there are a few steps every company should follow. Below I’ve listed the activities the marketing department should take in order to reduce wastage in time, budget and energies to finally unleash the potential of their data. Starting from the organization’s high-level objectives the marketing team should:
- Define operational needs. What data do you need to support high level marketing objectives?
- Analyse the as-is scenario. What resources are available? This covers team availability, existing data sources and data flows. Are you working efficiently or are you expending too much energy on low value activities?
- Define the to-be scenario. What resources are needed to support operational needs? Again, this includes human resources, processes, data availability, … Tip: since marketing budgets can be a constraint, organizations should try to eliminate redundant activities before thinking about automating tasks. You can try to apply Pareto’s Law to focus only on effective activities, Neil Patel provided a couple of great examples here.
- Take action! At this stage you can perform all the required tasks to move to move from the as-is to the to-be scenario. This may include a review of your processes, a change in how you collect and/or manage data or even the introduction of additional technology to better support your team.
- Measure. Are you improving your performances? As Peter Drucker once said If you can’t measure it, you can’t improve it; and so, it’s best practice to assess how your decisions and the changes you made actually helped you in achieving the desired results or not.
Conclusion
If you’re a club or a sports organization you probably have recognized several aspects discussed in this article. Hopefully, your best practice processes in data quality management are already in place, and you can describe yourself as a “data-driven company”. If this is not the case, and you need some support in assessing your needs and planning your next steps to unleash your data’s potential, please feel free to get in touch.
If you want to learn to use data to make business decisions and talk about data confidently, sign up to discover here about our eLearning course Winning with Data.
