- Interactive Rewards
Why you should clean your database every quarter (at least)
Q: What’s the similarity between cars and data?
A: They should both be cleaned every two-to-three months!
A recent report by Intelligent Car Leasing revealed that over a fifth of Brits never clean their cars, resulting in the interiors becoming a breeding ground for bacteria (not great in the middle of a global pandemic) and the exteriors becoming a safety hazard due to dirty windows and mirrors. Not to mention the fact that dirt can wear away paint leading to rust and eventual structural breakdown!
The benefits of regularly cleaning your car are clear and the people who do have a regular cleaning routine will go the whole hog – inside and out – rather than taking the bitty approach of windows one day and hoovering the interior the next. Like your car, data is another entity that should be cleaned on (at least) a quarterly basis! But many organizations take the piecemeal approach to cleaning their database.
When asked, most marketers would love the ability to clean their entire customer database on a regular basis. Not only would this help them to demonstrate their commitment and adherence to GDPR, but it also means superior targeting and better decision making. But the reality is that few do. The majority of marketers do it on a campaign-by-campaign basis due to cost.
This campaign-by-campaign approach means that when executing marketing campaigns, the target audience is defined first and then people within the database that fit the profile are selected. Then, and only then, is the selected data cleaned. The rest of the data is ignored. However, this method means that marketers run the very real risk of missing key prospects, reducing ROI, and leaving money on the table. Our estimations suggest that up to 5% of prospects fail to be selected because the data is not cleaned before the selection process.
For instance: Prospect A was added to the database three years ago when he was 28. At the time, he had a girlfriend but lived in a shared house with roommates. So far, he has never been selected for a mailing campaign because he doesn’t fit the profile. However, in the intervening years he has had a promotion, changed jobs, got married, had a baby, and moved to a new house. He is now a very attractive prospect, but he gets ignored because it was never flagged that he moved. This happens a lot!
Ironically, the cost of cleaning a database upfront would be more than covered by the incremental revenue gained by targeting these missed prospects. When you do the math, the ROI is clear to see. Let’s assume a database contains 500,000 records and 30% of these records are out of date. At a cost of 20 cents per flag, that would result in a cost £30,000 to clean the entire database. So, this means that the 5% of people that weren’t selected in the previous campaign-by-campaign cleansing model would only need to spend £4 each to cover the costs of the suppression. But regardless of this, for most businesses the risk/return profile is still too uncertain…
Until now! The data hygiene sector is changing. Instead of paying on a per record basis, it is now possible to pay for data cleaning on a subscription basis where the cost is calculated on an average matching basis (for example, through www.clean-contacts.co.uk). This means that brands would not only make a saving on their hygiene bills, but also bolster the bottom line simultaneously. It’s a win-win!
Furthermore, data hygiene is no longer a nice to have – in today’s analytically driven, AI world it is a necessity for effective data analysis. It is estimated that the analytics market will be worth a staggering $21 billion in just three years’ time, growing at a CAGR of 44%. However, it is well known that algorithms and analytics are only as strong as the data upon which they are built. So, it’s unsurprising that data scientists spend 60% of their time cleaning and preparing data for analysis. However, 57% of data scientists view this as the least enjoyable part of their job.
The bottom line is that if the data is flawed then the resulting model will be biased. At best this will result in poor targeting and wasted marketing budget, but at worst could have severe repercussions in terms of bias and unfairly targeting vulnerable customers, such as the recently bereaved. The argument for cleaning your entire database is much stronger than it has ever been and hopefully through the course of this blog we have demonstrated why the natural order needs a reshuffle!
This article was first published by MAPP. Permission to use has been granted by the publisher.