Is Bad Data failing your AI-Analytics investments- 5 things to know

We are at the cusp of Fourth Industrial Revolution...a digitally powered world where the line between Physical, Digital and Biological spheres are blurring. Each day we create 2.5 quintillion bytes of data spread across the internet, social media, communication channel and devices. This huge amount of data is like an ocean with hidden treasure...but only for those who can connect the dots and stitch a story. And for those who struggle...there is AI-Analytics !

But what makes an AI-Analytics solution work? Of all the elements in the equation, most important one is Quality of your Data. The accuracy with which your AI-Analytics platform is expected to generate insights is directly proportional to quality of data you feed. The lack of focus on data affects organization at all levels. As a Demand-Gen Marketer and Tech professional, I feel this frustration everyday. My wish-list for AI in Marketing is long and there are solutions in the market today that provide some excellent use cases. Will simple investment on these Technologies solve my pain. Short answer is NO. We must fix Data Quality issues first before embarking on an AI-Analytics journey.

Today's Marketing organization lives on data (both internal and external). Marketing data consumption ranges from Sales intelligence, Market intelligence, Customer insights, Channel performance to name a few. From customer segmentation to campaign attribution, the importance of structured data cannot be ignored. Global marketing organizations face the most pain from bad data due to siloed processes, definitions and reporting structure. The result of this inconsistency is emergence of a Data Archipelago  !

Have you wondered why we still have meetings where MS-Excel and pivot tables come to rescue? Why inspite of investing in intelligent platforms people still go back to pulling data in spreadsheets? The simple answer is systems are not communicating and hence users have to rely on excels and vlookup skills. Without proper Data Harmonization at Enterprise level, systems are just like teams without a common language. So where do we go from here? 5 things to get the data puzzle right:

1. Accepting Data Problem. The first step is to accept that there is a Data problem in the Organization. This has to be acknowledged by the Top execs who should support this initiative and are willing to take risks.

2. Ask Bold Questions . Ask questions on things that are assumed to be perfect in the Organization. It may be re-looking at your age old master data structure preserved in some sacred legacy platform. It can be re-visiting your Account structure which is considered perfect, it can also relate to your Product groups or legacy databases that have not been functional for last 2-5 years.

3. Kick-off Enterprise Data Assessment. Get your Business SMEs and IT Data experts together. Ask them to classify your Enterprise data into 3 segments: Business Critical (Your business cannot run without this data), Business Relevant (Your business will function better with this data), Non-relevant (Data that does not add much value in your Business). Empower Business SMEs to identify and take decision on keeping or discarding non-relevant data segment. IT to focus on harmonizing Business Critical/ Relevant data.

4. Establish Data Governance. This includes implementing consistent data rules across platforms (upload/ extraction/ de-duplication/ merge), identifying internal Data Champions at Enterprise level as well as Intra-team level, defining Audit process and schedule. This becomes building blocks for your Data Management CoE for future .

5. Finally, define Data Owners. So who should owns data in your company? There is no single ownership of data. DO NOT make the mistake of assuming IT as the owner for your data. This is the classic mistake most companies do. Data has to be owned by each person who has access to it and can create, edit, delete data. Data management CoE will ensure data rules are followed and regular audits are performed. But, at the end it is your end users who have to own it. Easier said than done, this has to be reinforced via continuous Training emphasizing benefits of good data and impacts bad data for the company.

Quoting Craig Mundie (Senior Advisor to CEO at Microsoft), "Data are becoming the new raw material of business.” Good data quality is no more a choice for businesses, it is a must. And it can be achieved with well planned initiatives and governance.

Subhayan is a Marketing Technology professional. In his current role he is helping Cognizant develop an integrated Global MarTech strategy connecting marketing efforts across channels, increasing engagement and generating demand.