Adland is full of conversations about data – what to collect and how to collect it, how to draw insights from it and how to craft compelling stories that resonate with audiences around it. Jonathan Cotton finds turning it all into real commercial runs on the board is harder than it sounds. So what do data-driven marketers – and their collaborators – too often get wrong?
“Being data-driven means you can focus on the things that are going to really make a difference to your top or bottom line, or to your customer experience,” says Nathalie Morris, CEO of marketing intelligence company Qrious.
Morris says that fundamentally, data-centric marketing is an attempt to replace guesswork, gut instinct and doubt – the wheelhouse of much brand-focused storytelling – with Real McCoy empirical evidence.
“Data can help you to truly understand your target market and know with certainty what your ideal customer looks like, as opposed to making assumptions.”
That’s all well and good, but as anyone who’s tried it knows – taking data and transforming it into solutions that achieve both business and creative goals is no mean feat. How does data – in all its abstract glory – fit into the commercial picture for most companies?
“The first place to start is with your company strategy and objectives,” says Morris. “What are you trying to achieve?”
That can be a broad subject in itself, but essentially comes down to two finding out how data can inform two core areas of the business, says Morris.
“The first question is ‘what data do you need for financial and business performance i.e. to make effective decisions around revenue, costs, or business operations?’ And the second is ‘what data do you need to provide unbeatable customer experiences, for example, if you want to provide a seamless customer journey from online to instore, or a frictionless call to a call centre?’”
That second part – understanding the customer to build better relationships – is where business processes, customer-centricity and the marketing department’s mandate starts to overlap.
“Marketing has always been about relationships and storytelling,” says Matt West, partner at design/creative agency EightyOne, “capturing how your customers relate to you, then figuring out how to find more of the people who like you and what you stand for”.
“If you capture how and when they relate to you, you’ve got somewhere to start from. That includes responses, transactions, interactions. But just collecting it because it’s there is pointless.”
“Working out a strategy, then collecting what you need to execute against is the best plan. We’d suggest organisations start simple. Understanding who your customer actually is and how they interact with you is a great start.”
Simply put, business strategy comes first. When you’ve established that, it becomes easier to build a data strategy, specifically.
“The ‘why’ will inform the ‘what’,” says Morris. “Once you know that, you can determine factors such as where the data will come from, where and how you’ll store it, how it will be accessed, what systems it needs to integrate with, how you’ll maintain privacy and security. From there, you can develop a roadmap for implementation.”
So far, so good. But if it was that easy, everybody would be doing it. Just what is it that disrupts so many well-laid data collection plans?
When it comes to doing business in 2019, responsible data management – or more specifically security and privacy – are the issues du jour.
With the Eurpoean Union’s GDPR, Australia’s deeply troubled data privacy legislation and New Zealand’s rapidly evolving data compliance requirements, establishing trust – and living up to it – is the challenge of the moment.
“Many customers have concerns about sharing data if there is any doubt about how that data will be used,” says Justin Flitter, CMO at data strategy company Data Insight.
So they should. Data scandals are now everyday news and customers are increasingly cognisant of their rights when it comes to user data use and abuse.
“The most important consideration in the beginning is ‘is this information I am sharing relevant to the experience or service I am expecting the company to deliver?’” says Flitter.
Take data collection based on an online form as an example. If questions about sex, ethnicity or religion aren’t directly related the offer at hand people will be reluctant to provide that detail – and subsequent attempts to establish trust may be rejected. A worst case scenario is if it all goes wrong and you yourself in breach of increasingly strict and complex data collection regulation.
“In all cases it’s important to explain to customers what data you’re collecting and how it will be used to show a fair value exchange.”
West puts it in simpler terms: “Permissions and privacy are the biggies and keeping on top of the legislation can be a job in itself.”
“Just make sure you have permission and purge anything you’re unsure about. Safety first.”
Behind every privacy breach headline is a company ignorant – sometimes willfully so – to their obligations under the law. But even beyond the clear security risks, poorly managed data provides a slew of other headaches – including both accessible and usability issues.
“The biggest challenge we see among our clients is data silos,” says Morris, “data in multiple separate systems and all in different formats”.
Inaccessible data – or more specifically, collected information that is only accessible to a certain group or area of the business – is the dreamkiller for those looking to unlock data-driven marketing potential. Simply put, it you can’t access it, it can’t be used to create insights into the business – or to support marketing activity.
“We recently worked with a client to implement a new marketing automation ecosystem,” says Morris. “To deliver personalised marketing programmes the platform needed a view of customer profiles – i.e. who the customer is, what products and services they have, their contact history, etcetera.”
“But this data was stored in multiple systems, and those systems had different definitions for a ‘customer’ – in some it was a person, in some it was an account. We also had to determine how to deal with multiple people or multiple services on one account.”
For Qrious, that meant that the biggest obstacle wasn’t the platform implementation itself, it was accessing and preparing the data that feeds that platform – a common problem for companies just starting to move towards more data-centric business strategies.
Similarly, error-ridden databases of ‘rogue’ data can come in myriad forms – incorrect details such as birthdates and name spellings, improperly formated material, and various issues arising from the merging of disparate data sets – but all have the potential to quickly undermine the efforts of the unwary would-be data marketer.
“As you become more sophisticated with your use of data and start developing predictive analytics, such as churn or propensity models, incomplete or dirty data can have a major impact on the accuracy and effectiveness of these models,” says Morris.
And make no mistake – the threat is real. Without ‘clean’ data, you’ll be delivering neither relevant customer touchpoints nor making accurate, data-driven decisions. And while there are methods of fixing flawed data sets, the best approach is, by far, to do it right the first time, says Flitter.
“Increasingly there are data cleaning tools analysts can use to process datasets, but the general best practice is to format datasets from the start to reduce formatting and structure inconsistencies,” he says.
“The challenge in collecting data is ensuring clean, consistent data is collected that can be used for analytical purposes. Businesses need to ensure data collected is verified or limited to a set range of options where possible and essential data is mandated to avoid gaps.”
It’s an all too common scenario for New Zealand business, says West.
“I’ve seen plenty of data lakes with the same effluent levels as a South Canterbury dairy farm,” he says.
“We say ‘start with good data’ which means ‘start with the right information’. Figure out what’s important, and what isn’t because ‘good’ doesn’t mean everything.”
It all adds up to a bigger, more time consuming – and potentially costly – process than some expect. To get the most from any marketing activity, cross-company support is a must-have.
“Improving the cleanliness and completeness of your data requires you – as an organisation – to look at the whole data ecosystem from end-to-end, and identify areas or processes that need improving,” says Morris.
Indeed, while the management of data itself can be a complex task, it’s usually not the technical details that trip marketers up. Rather, it’s a failure of the business as a whole to appreciate the scope of the undertaking.
But that buy-in is crucial to achieve real data-driven decision-making – in the marketing department and across the business.
“Data-focused companies share trusted data across the organisation,” insists Flitter. “They act on fact, not theory, making informed decisions faster. This empowers people at all levels to receive and understand data and the business insights to make data-driven decisions.”
“This isn’t about sharing spreadsheets with screeds of tables, it’s about surfacing the most relevant insights people need to make decisions faster.”
The more you know about someone, the better you can serve that customer and the better you can tailor the creative, says West. “Data can help tell you the how, the where, the when and the why. We can anticipate life stage, understand mood and figure out what would be most useful next, but the biggest challenge is a reluctance to evolve or change from inside an organisation.”
“Plenty of people are willing to change of course, and they’ll be the ones to kick goals in the future.”
