Let’s be honest, talk of building a data quality strategy doesn’t match the seductive spiel of the sales agent promoting the latest cure-all platform. But it’s boringly important. Data decays faster than you’d expect and corrupted datasets can multiply at speed.
The result? Bad decision-making, ill-informed marketers and regulatory risk.
All the constituent parts of a mature data-informed organisation are outlined below. If you want to become a masterful wrangler of data then it’s worth taking the time to understand these concepts and their practical application.
Why? Because if you get to grips with all this, then you’ll be able to:
- Draw the most value out of your data
- Ensure you comply with data protection regulations
- Run a more efficient and productive organisation
- Make the wisest investments in data and marketing technologies
In short, a data quality strategy (which is part of your broader data strategy) provides the bedrock for ongoing success as a data-driven organisation.
Where should you start with a data quality strategy?
We explored some of these themes in a whitepaper with data quality specialists, REaD Group. In particular we looked at the issue of data quality, the scale of the ‘bad data’ problem and practical steps for ensuring data is clean and ‘dependable’.
Your data strategy is the most importance place to start. This is reinforced by the Insights2020 study which discovered that:
- 67% of execs at over-performing firms were skilled at linking disparate data sources
- 61% of over-performers have ‘Insights’ involved in all key areas of planning
- 71% of over-performing firms combine analytical and creative thinking.
There are many facets to the data strategy (and you can read our in-depth guide on how to build a data strategy here), but two key areas are the:
‘Data Architecture’ and ‘Information Architecture’.
Both play a key role in ensuring that your data is primed for optimal use.
What is ‘Data Architecture’ and ‘Information Architecture’?
According to the Harvard Business Review a company’s Data Architecture describes: “how data is collected, stored, transformed, distributed and consumed”. Its underlying purpose is to ensure your data is ready to meet the needs of the business and is fit for processing.
It’s not just a technical consideration, it’s a business decision. After all, new business models and entirely new ways of working are driven by data and information. The Data Architecture includes an organisation’s governance and compliance frameworks and it provides the foundation for the ‘Information Architecture’ which is the process of transforming raw data (names, addresses, metrics) into useful and actionable information.
Or as David Loshin and Charles Roe from DATAVERSITY summarise it:
“Information is data put into action.”
The Data Architecture should always be a Single Source of Truth – it’s either right or wrong. However, your Information Architecture could display multiple truths depending on how it’s been processed or consumed. They may or may not be accurate and part of your data literacy training should make it clear that all processing innately has bias and can often show what the creator was “looking for”.
Reliable data starts with a robust Data Architecture
While it’s clear that all the constituent parts in the Global Marketing Alliance model are important, some, due to the dependency others have on its outputs, are more critical than others. Having a robust Data Architecture is one of these. It’s essential for delivering your data strategy and for enabling almost every other part of the model to work.
As the well known expression goes: “garbage in = garbage out” and it’s the role of your Data Architecture to deliver this.
Unfortunately, it is often poorly resourced compared to other areas which have more instinctive appeal to the business. After all, visual dashboards are more appealing than business models.
The REaD Group framework for Data Architecture
REaD Group identifies five core pillars which, when combined, enable your data for optimal use within the organisation. This high quality, reliable data feed can then be passed into your Information Architecture layer for processing with a high degree of certainty that it is accurate and compliant.
1. Data Collection
How, where, when and what data is collected.
2. Data Storage
Where data is stored and ensuring a single source of truth.
3. Data Quality
How you clean and append data to make it ready for processing.
4. Data Governance and Compliance
The rules that govern your data systems including your legal obligations.
5. Data Distribution
How you make data available to where it’s needed, when it’s needed.
How should you define data quality?
Quality data is useful data. It must be consistent and unambiguous. Problems with data quality are often the result of database merges or systems/cloud integration processes in which data fields that should be compatible are not due to schema or format inconsistencies.
Data quality deteriorates unless it’s cleaned and updated on an ongoing basis. In a similar way to radioactive material, data has a half- life. Understanding your data’s half-life helps frame your approach to data cleansing. It also impacts a number of other areas of your data strategy.
Your data half-life is:
“The time it would take, if there was no remedial action, for 50% of your data to be inaccurate on a critical data field.”
A range of factors will dictate the speed of your data’s decay:
Sector: For example, food retail vs banking.
Job roles: Some roles turnover very quickly, some less so. The average tenure of a marketing manager is 24 months whereas a CFO is more than 4 years.
Life stage: Under 30’s are much more likely to move house, change job, get married than someone in their 50’s.
Depth of data: The more critical data you hold, the more opportunity there is for it to decay.
However, this doesn’t include other potential sources of data inaccuracy, such as the proliferation of incorrect form entries or duplicate contacts.
Questions to consider when defining your companies clean data state:
- What is your definition of ‘is it clean now?’?
- What is the margin of error that is allowable for your business, if any?
- Do you know where all your data is and who’s responsible for it?
- Do you have an actionable data maturity model in place?
- How you answer these questions should provide a clear sense of where you need to start.
“To genuinely understand the value of data, business leaders must ask these questions, within the context of their environment and business objectives, and use the answers to develop a data strategy to drive responsible and profitable decisions.
“Once tangible value can be attributed to data, then the business case to justify the time and investment for maintaining the accuracy and quality of that data, on an ongoing basis, should write itself.”
– Firas Khnaisser, Head of Decisioning at Standard Life and Chairman at DMA Scotland
How to create a data quality culture
“Data and analytics leaders need to understand the business priorities and challenges of their organisation. Only then will they be in the right position to create compelling business cases that connect data quality improvement with key business priorities,” Ted Friedman, vice president and analyst at Gartner.
Your data strategy needs to feed off of the business strategy – and vice versa. Linking the two highlights data’s strategic importance to the wider organisation.
It is surprising how many businesses do not have a formal data strategy in place. They collect and hold unprecedented volumes of data – and might acknowledge that data is an important business asset (even though they can’t value it). But they do not ask themselves what is the value of this data? Why do they collect it? Why do they maintain it? And are they using it to deliver value to the business?
Perhaps counter-intuitively, getting clean and staying clean is more to do with culture and less to do with process and technology.
It’s easy to perform a data “spring clean” as a one-off project. But this approach simply resets your data half-life and fails to provide the long term solution that data-driven organisations require. Fast forward just a few months and you’ll likely be back to the start again.
Creating a “Data Quality Culture” is therefore essential and it must be initiated at the top of the organisation. It is not just a matter of implementing strong validation checks on input screens because no matter how strong these checks are, they can often still be circumvented by the users.
Industry pioneer, Ralph Kimball outlined the general actions you should consider implementing to help build a data quality culture:
1. Declare a high level commitment to a data quality culture
2. Drive process re-engineering at the executive level
3. Invest to improve the data entry environment
4. Invest to improve application integration
5. Invest to change how processes work
6. Promote end-to-end team awareness
7. Promote interdepartmental cooperation
8. Publicly celebrate data quality excellence
9. Continuously measure and improve data quality
Ensure data drives you in the right direction
As we can see, data quality is an ongoing pursuit. But once the fundamental processes are put in place, including ongoing (or scheduled) data cleansing procedures, you can be assured that your data is pointing you in the right direction.
A dedicated data quality strategy should be a key part of your wider data strategy which in turn draws on your business strategy.
Want to know more? Download our free whitepaper produced in collaboration with data quality specialists, REaD Group: Are you addicted to bad data? Get clean and stay clean