To many, marketing now is all about data.
In our 2016 Quarterly Digital Intelligence Briefing, more than half (53%) of marketers surveyed said ‘data-driven marketing’ was their top priority this year.
Nine out of ten put it in their first three, more than any other topic.
But what are marketers actually doing with their data?
What tips can professionals give for those who may be just starting out with data-driven marketing?
To find out, Econsultancy recently held roundtable discussions at our fifth annual Digital Cream Sydney.
There, client-side marketers from across the industry discussed trends, best practices, and the issues they are currently facing.
The roundtables were moderated by subject matter experts from the industry. Participants brought their own experiences, questions, and challenges to the table for open discussion.
Here are the highlights from the discussion at the Data Driven Marketing & Marketing Attribution Management table.
1. Use personas and customer journey mapping for attribution modeling
We now live in an omnichannel world. People often use the web, social media, mobile, and search before buying something.
How can marketers determine the right amount to invest in each channel?
Participants agreed that doing so, also known as attribution modeling, is one of the toughest tasks marketers now face.
Figuring out which channels drive awareness, which help with research, and which lead to conversions is not easy – even with all the data in the world.
While attendees admitted that there is ‘no silver bullet’ for determining the right model, delegates suggested that using customer experience data can help.
They said that creating audience personas and then mapping each customer journey can provide insight into the path-to-purchase for different customers.
This can then provide the foundation for the elusive attribution model which helps marketers allocate their spending for optimal results.
2. Avoid using personas for more granular data-driven marketing
While the customer-centric approach may work for modeling attribution, delegates agreed that personas and customer journey maps were not so useful when doing more personalised data-driven marketing.
That is, when buying programmatic media or providing on-site personalisation, broad segments and models do not help.
Instead, attendees stated that marketers should use an individual’s behavior to deliver relevant ads and personalised content.
What a person has viewed or purchased previously is much more likely to attract their attention in the future than something which fits a particular persona, one participant argued.
3. Look at your data when optimizing
Another dilemma marketers often face is how to optimize their website and ad buying based on outside trends.
Recently, there have been many charts showing that mobile traffic is outpacing web traffic. Does this mean that marketers should go ‘mobile first’?
Not at all said the delegates. While it is useful to be aware of the trends in mobile, video, and messaging, marketers should prioritise their own customers’ behaviours to help form strategies.
As an example, at one table on the day, there were some marketers who said that mobile usage was plateauing while others said that tablet traffic is becoming increasingly important to them.
So, the recommendation is that marketers should first keep a close eye on the trends in their own data before making any drastic changes as a result of industry reports.
4. Use data for more than just conversions
Marketers these days are typically required to produce data to justify their budget.
Metrics such as cost-per-acquisition (CPA) and return on ad spend (ROAS) are commonly used by the business to gauge performance.
Because of the need to demonstrate that marketing spend matters to the business, attendees agreed that most of the effort spent on marketing attribution and data-driven marketing is used to lower customer acquisition costs.
However, delegates also agreed that we now have the data to do much more. Data should also be used, they argued, to improve customer retention and loyalty.
Doing so will, in turn, increase the lifetime value of customers and improve the bottom line, albeit in a less direct way.
Marketers should, therefore, look for opportunities to use data for customer experience and resist the tendency to look for the immediate gratification of a lower CPA.
5. The best third-party data is from sites where users log in
While marketers tend to have a good handle on the data from their own sites (first-party data), many are still wondering about the value of data from other sites (third-party data).
This concern was made apparent because, when asked, only around 10-15% of marketers at the tables admitted using a data management platform (DMP) as a ‘single source of truth’ about their customers.
The reasons for hesitating are well-founded. Many third-party data services guess at aspects of users’ identities from the sites they visit or activities they have done in the distant past.
Attendees asserted, however, that sites which require users to log in can provide much higher-quality third-party data.
Specifically, Google and Facebook can both link extensive browsing and posting behaviour to a particular person.
For this reason, delegates said that such sites do offer third-party data worth using for advertising and analytics.
Interestingly, one participant noted, both Google and Facebook are also starting to offer data which allows brands to track consumers offline.
That is, they will know whether someone has entered a particular location (e.g. a store) after viewing an ad on their platform.
6. Aim to make small changes with insights from data
While most of the day’s discussions were positive, one negative aspect of data-driven marketing emerged.
Even with insights from data, delegates admitted that it was rare that recommendations based on data were actually implemented.
Data was more likely, they said, to be used for retrospective reporting and business-oriented statistics.
One way around this, one participant suggested, is to adopt a more ‘agile’ way of working.
What this means is that marketing teams should avoid gathering vast amounts of data in an attempt to influence strategic decisions.
Instead, marketers should use insights to drive incremental changes on a frequent, tactical basis.
In this way, the ‘agile’ approach will change an organisation’s approach to marketing iteratively over time and have a much higher likelihood of succeeding.
7. The biggest hurdle? Finding the right people.
In previous years, marketers have lamented about quality of marketing technology and the difficulty of obtaining data to drive marketing strategy.
While these are still concerns, delegates this year said that their biggest challenge was finding the right people to drive data-driven marketing initiatives.
Attendees agreed that that finding people who could interpret data both technically and commercially was really hard. Additionally, these people are critical for getting insights out of data.
Newly-hired data scientists are often too technical and abstracted from the operational business to help. Experienced marketers, though familiar with the business, often lack the statistical modeling skills to extract new insights from data.
One suggested approach is for marketing teams to recruit analysts with business acumen and data crunching skills.
But in lieu of staffing up with the right people, participants felt that marketers could also take a more active role in interrogating the data themselves for insight.
A word of thanks…
Econsultancy would like to thank all of the marketers who participated on the day and especially our Data Driven Marketing & Marketing Attribution Management table moderators, Beaudon McLaren, APJ Ecommerce Manager at Symantec and Ashley Friedlein, President of Centaur Marketing & Founder of Econsultancy.
We hope to see you all at future Sydney Econsultancy events!