Good UX needs big data: how Master Data Management (MDM) can syphon your data-lake to fuel AI
Despite the rise of artificial intelligence (AI), robotic software is yet to be able to think creatively or perform quality assurance on themselves, which can negatively affect a user’s interaction with your product or service.
Good data in means a good user experience out, so if you want to be ready for AI and robotic process automation, you need to make sure your spreadsheets are in order.
Feed your robots
Let’s say a customer logged onto your site and started up your chatbot:
Hello, Smith, John. How can I help you today?
Well, this is Mrs Smith. His wife.
Hello Mrs Smith, John. How can I help you today?
I was just checking up on my outstanding order?
Is it the one with the reference ########?
That’s not a number
Is it being delivered to 45 Stansted Road?
No, that’s my old address
It seems like we’re having trouble. Can I get someone to call you back on ##HREF###?
A human seeing ##HREF### would assume something had gone wrong; a robot program just looks up the field in a database and churns out the contents without considering whether it’s right.
This means that whatever you feed into a robot is what you get out at the other end. So, you better make sure you’re feeding it good data.
What does ‘good’ data mean?
So what is good data? It doesn’t just mean accurate information, though that is a great start. Outdated information from old systems and files is just a spanner you don’t want to throw into the works of your robot.
Still, even valid data can be an issue if it’s in the wrong format. Searching the records for Mr Smith, John isn’t going to get you very far if the information is saved as J. I. Smith, Esq. Likewise, you won’t be able to auto-dial internationally if your phone numbers are saved as the standard “07XXX XXX XXX” rather than “+447XXXXXXXXX”, meaning your call centre staff have to manually edit it each time, reducing their efficiency.
Even if your data is accurate and in the right format, that won’t matter if it’s in the wrong place. Forcing your robot to trawl through myriad different files to find the particular location a specific piece of information is stored will make it slow and take up bandwidth. For an efficient robot, you need an efficient filing system in one, single location.
Master data management (MDM)
To leverage all this big data, companies are taking every piece of information that flows through their systems and funnelling it into one place as a big pool of unformatted, raw data. This ‘data lake’ doesn’t discriminate, meaning everything is stored and ready to use. A 2017 survey by Big Data LDN saw 93% of respondents say they were planning to move to a data lake by around the present day.
Master data management, or MDM, is the discipline of taking the relevant data from your data lake and formatting it into a single master data file with a consistent format and system of reference. This can also be broken down into ‘data warehouses’ for different outcomes.
As an industry, MDM will be worth $22 billion by 2023, according to Markets and Markets; and it’s easy to see why.
One database to rule them all
Instead of telling your robot to go look at your insurance customer database, under the renewal system, in the correct country and search for John Smith; you can tell it to just search for John Smith, narrow down those results by country and then access all the data for that customer.
Maybe Mr Smith changed his address on one account, but not the other. No problem; the new address is on his main record, rather than only one system.
Perhaps he queries the chatbot about insurance, but then wants to check on the status of his holiday booking. No worries; both accounts are linked in the same record.
Just as it would be easier for you to find the right info under this system, so is it quicker for your bot. Where your bot works quickly and effectively, it can deal with more customer interactions and save you time and money.
Your company likely has a Twitter account. Does your social media team archive their tweets? Do you analyse responses? If you keep that data in a lake, then synergies become possible.
Using automation you could, for example, identify a tweet of complaint from a customer and log it against their account. Your customer relations team could then pick this up and contact the customer directly to apologise and rectify the problem before it causes them to cancel their policy with you.
In 2016, Twitter claimed that over 70% of people who complained to a company by tweet expected a response within an hour. Do you think they’d prefer, “DM me your account details and I’ll look into it for you” or “Hi Steve. I’ve already gone into your account and fixed that”?
That’s just the start…
Big data means big prizes
Perhaps your robot could identify that a customer has a health insurance policy for a 17-year-old and car insurance with you. Could your team contact them to offer to add a learner onto their insurance at a discount?
Maybe they’ve booked a holiday with you for after lockdown. Can you offer them a bolt-on for travel insurance? Have they recently made a private medical claim for care for a particular condition that you can offer tailored insurance for?
Say they recently accessed your site from an iPhone for the first time. Could you email them a discount code for iPods and a spare charging cable?
What if one of your sales team is speaking to the CFO of a potential client, but another is speaking to the COO? Storing their contacts and leads in a managed data lake could lead to that being flagged with both of them before a faux pas occurs.
The benefits can even be as simple and effective as your chatbot knowing whether your customer prefers to be called John, Jonathan or Mr Smith.
Benevolent Big Brother
This may all sound a little bit Big Brother-ish, but the key is that you’re using this detail to offer customers products they want and may even need. This data is fueling automation, which will offer a better service and a better user experience. Bearing in mind that 86% of customers are willing to pay more for a better service, according to SuperOffice, can you afford not to think about it?
Solid MDM makes automation not only worthwhile, but feasible in the first place. It’s the difference between companies that users see as efficient and competent, and the ones that leave their customers in a call queue waiting to unsubscribe.
Get your remote lab running and delivering UX insights today!
UserZoom’s remote usability testing solution, Validately is a quick and simple way to start listening to your users in moderated or unmoderated interviews.
You can also collaborate with stakeholders, automate scheduling, curate reports – all completely remote, so you can iterate from anywhere in the world and continue making great digital experiences.
Try it for free, no credit card required
Neil Sheppard has been a UX copywriter and content strategist for nearly a decade. Starting out as a pharmaceutical stock markets journalist, Neil quickly moved into digital copywriting, managing a team optimising product content for a busy commercial website. Nowadays, Neil helps companies create easy-to-use internal websites and digital employee manuals that make complex processes simple for everyone from CEOs to service desk agents.