Applying short-term data to long-term thinking: A beginner’s guide to data-led planning

In recent years, our industry has become increasingly obsessed with using data to optimise one-to-one targeting and performance-based marketing.

As the negative implications of too much emphasis on short-term conversions have become apparent, however, there is also an opportunity to think more broadly and ask ourselves:

How can we use this wealth of ‘new’ data to enhance strategic thinking and liberate creativity across the entire communications approach?

Media and creative strategy have historically been founded on linear, static research methodologies with one fundamental issue: they’re based on claimed beliefs and behaviours, with no guarantee they’re a true reflection of how consumers actually behave.

Instead, in the same way that digital targeting is fuelled by behavioural data, so too is the wider planning process, meaning our thinking can now be founded on demonstrable truths about human behaviour.

This leads to a fundamental shift in our approach to gathering insights, although the overarching questions we must ask ourselves remain unchanged:

Ambition: What is the business and brand ambition we are trying to deliver?

Audience: Who should we be targeting in order to achieve these objectives?

Brand: What is the audience’s relationship with our brand and broader category?

Barrier: What are the barriers we need to overcome in order to deliver growth?

Messaging: What messaging and content will influence our audience?

Media: When and where is our audience going to be most receptive to our message?

Culture: How can we tap into culturally-relevant moments to amplify our activity?

In the following sections, we’ll explore ways to embrace this opportunity: outlining the key steps to take; the differing characteristics of 1st, 2nd and 3rd party data; and providing examples of putting data-led insights into action.

Sharpening your axe

The defining characteristic of behavioural data sources is the almost infinite number of data points available. That’s because every single digital interaction is generating a data point of some description. The key challenge isn’t finding the data – it’s knowing what to do with it.

To avoid choice paralysis, it’s imperative to have a clear plan of attack. As an anonymous woodsman once said, “if I had five minutes to chop down a tree, or pay with my life, I’d spend three of those minutes sharpening my axe”.

To ‘sharpen your axe’ for successful data-led planning, there are eight key steps:

1.    Start with the strategy: Be clear on which overarching question you are looking to answer. This should ideally be one of your overarching strategic questions.

2.    Define detailed questions: Identify more granular questions that could potentially help you answer the broader strategic questions.

3.    Establish a hypothesis: Use your existing knowledge of the category and consumers to establish a theory about the answer to that question, which you can then test through the use of behavioural data.

4.    Identify data sources: Conduct a data audit to determine data sources that could help you test these hypotheses.

5.    Consider a proxy: If there isn’t a data point available that answers your specific question, there may be other data sources that can provide an indirect way of testing your hypothesis.

6.    Combine the data: Look at opportunities to cross-reference data from different sources to identify interesting correlations or contradictions. In the absence of attribution modelling, these can often unlock surprising insights.

7.    Understand why: To unlock the insight, that deep human truth that will fuel your strategy and ideas, you must ask why consumers are behaving in this way? It’s in understanding the ‘why’ that you’ll be able to make the leap into actionable insights.

8.    Turn insight into action: Take the various insights and filter them down to those which are truly meaningful, by defining the specific action the brand should take as a result of that insight. If there are no clear implications, then drop it from your thinking.

Strategy is as much about what you choose not to do, so get focused, fast.

1ST PARTY DATA

It’s easy to develop a preconceived notion of how and why customers engage with your brand’s products and services, whether that’s based on your own personal experiences or an idealized vision of the role your brand plays in people’s lives.

This is why your brand’s own data can be so powerful, because it provides unencumbered insight into how customers interact with your products and owned media platforms, helping to unlock deeper truths into the actual human need your product or service fulfils.

Data from the brand’s own website is typically one of the richest sources of insights, encompassing both customers and engaged prospects, while data from loyalty program members provides insight into the audiences most likely to heavily engage with your brand.

Before you begin, however, it’s crucial to sharpen your axe, by determining the specific questions you’d like to answer in order to inform your broader approach.

Example 1st Party Data questions

Ambition: What combination of products are people reviewing or purchasing on our website/app? What is the current volume of visitors to our home page versus specific product pages?

Audience: What’s the demographic profile of our customers e.g. age, gender, location? How does the demographic profile vary e.g. for different products, new vs repeat buyers?

Brand: How regularly are prospects and customers visiting or contacting the brand? What searches and questions are visitors submitting on our website/app?

Barriers: What products and pages have the highest bounce rate i.e. where visitors drop out? What is the proportion of repeat vs new visitors to the website and how has this changed?

Messaging: How does my engagement rate vary for different on-site messages? How does the performance of different messages vary by placement?

Media: When are people engaging with my owned media assets? How have the sources of referral traffic to my website evolved over the year(s)?

Culture: How does customer behaviour change during key cultural moments across the year? Have there been any unusual peaks of visitation not associated with marketing activity?

Applying 1st Party data

The options are almost endless, limited only by the data measurement systems you have in place, and your creativity in finding ways to slice and dice the data.

Looking at geolocation data from users of your mobile apps can help predict where customers live and work, which can then be matched with localised census data to provide rich demographic insight. Messaging and media – including out of home – can then be tailored to the unique needs of each region.

From a seasonality perspective, cross-referencing ecommerce search and sales data can indicate at what point in the year people are planning versus buying products and services. In the case of travel, the booking lead-time can vary significantly for different occasions, such as Christmas holiday travel versus long weekends, enabling you to adapt above-the-line messaging and flighting to ensure it is in sync with online behaviours.

1st party data can even be used to optimise individual TV program placements, by matching each spot with the volume of website interactions that were generated. Statistical analysis isolates the incremental contribution of each spot, maximising the effectiveness of any direct response TV activity.

Case Study: Booking.com

The team at Booking.com were analysing geolocation data from mobile bookings when they noticed something surprising: there was an unusually high proportion of searches being made along the freeways into and out of major cities around the world. Furthermore, these searches were for hotel bookings that very same day. In short, people were spontaneously packing their bags, jumping in the car and then hitting the road, all without having even booked a hotel. For this generation, ‘winging it’ was part of the fun.

From there, the campaign idea of “Wing it” was born. Booking.com embraced this simple yet powerful insight, with a campaign in the last days of summer that aimed to inspire all intrepid young travellers to ‘wing it’ before it was too late. Launching with a TVC and the hashtag #wingityeah, they directed travellers to an online hub, where Booking.com turned their jealousy-inducing travel pics into animated gifs and one giant viral mashup video.

By tapping into the brand’s own behavioural data insight, Booking.com was able to become known for making the most of the last days of summer. The campaign delivered 16.4M views and over 500K unique visitors to the #wingit hub, but most importantly contributed to a 48% increase in consideration.

2ND PARTY DATA

The application of 2nd party data allows the strategic planner to look beyond existing customers and engaged prospects, delving more deeply into the content needs, interests and opinions of the broader market.

Data from social platforms in particular provides insight into the demographic profiles, interests and content engagement of the followers of brand and competitor pages; while search data can improve understanding of interest in content, competitors and receptivity to specific messages.

In many instances 2nd Party data can also be used as a substitute for information the business lacks, such as ecommerce partners sharing purchase and post-purchase behaviours.

Campaign interactions can also provide a wealth of information about consumer receptivity to specific messaging, moments and mindsets, all of which can inform broader strategic thinking.

Example 2nd Party Data questions

Ambition: Which products are people reviewing or purchasing via our ecommerce or affiliate partners? How has search volume for my competitors evolved over the last five years, 12 months or 24 hours?

Audience: What is the demographic profile of people who follow my social pages and those of my competitors? What unrelated social pages and content are my brand’s followers most likely to share or like?

Brand: Which keywords do people most commonly enter in search engines when they’re researching my competitors or the category? What competing brands do my social followers also like?

Barrier: What negative comments are getting the most traction on our social pages? How are customers rating our products or services on affiliate and ecommerce platforms?

Messaging: What kind of content and formats are people most likely to engage with on our social platforms? What are people saying about our brand on our own social pages?

Media: Where are people currently getting their information from when they search keywords related to my category e.g. Wikipedia vs forums? What kind of content does my audience engage with on partner apps or websites?

Culture: What are the current rising search trends in my category? When are there specific peaks or dips in search or social interest for my category?

Applying 2nd Party data

Search data is a particularly rich vein of ‘nuggets’ to mine, such as using monthly search volumes for competitors in your category to estimate the share of search volume for each brand, which can be used as a substitute for active consideration or even market share if that data isn’t readily available.

Search volumes can also be filtered by day of week or time of day using Google Insights, meaning specific moments and mindsets can be inferred, while annual seasonal trends can inform your understanding of when demand-based messaging should be dialed up in your long and short-term flighting plan. 

Using Google Keyword Planner to look at the search keywords associated with your competitors can identify more strategic areas of opportunity, particularly when cross-referenced against your own brand search data.

The performance of different ad copy and offers in paid search campaigns can also be used to adapt messages and promotions in offline channels such as TV, with one insurer in the UK revising their on-air offer from 20% off car insurance to 10 weeks free based on search keyword performance. Business return increased significantly, despite ‘10 weeks free’ being worth marginally less at 19.2% off.

Geolocation data from partner apps adds another layer of real-world insight, such as using data from running and music streaming services like Strava and Spotify to identify popular running trails by city, targeting regular runners with tailored messages, music and outdoor media.

When overlaid with competitor store locations, mobile ad campaign data can go even further, identifying the behaviours of category buyers, including category duplication, time spent in-store and the catchment area for each store.

Case Study: Pepsodent White Now

Pepsodent White Now is a toothpaste that makes your teeth instantly whiter after just one brush, with the tagline that ‘a smile changes everything’. With their target audience of Millennials prone to ‘duck face’ selfie pouting, however, the brand needed to find a way to bring proof to the positioning.

Enter Tinder, the ubiquitous dating app for Millennials. Taking live data from 93,700 data points, the team compared profile pictures of smiling users versus unsmiling users, to arrive at a truly compelling insight: 70% of Tinder users saw an uplift in swipes right (an invitation to ‘make contact’) when flashing their pearly whites.

The team turned this insight into a fully integrated PR and media campaign, launching the first ever Tinder campaign in Sweden. Using creative executions which dramatised the impact of having a smile on dating success, the messaging also fed these live data insights directly back to the original study participants. With an average engagement rate 2.6 times higher than campaign benchmarks and an increase in brand consideration of 17%, the campaign proved that a smile really does change everything.

3RD PARTY DATA

While 2nd party data involves partners sharing their own 1st party data, the provision of 3rd party data comes from providers who access information from a multitude of external sources.

Similar to 2nd party data, it can be invaluable for providing wider category, consumer and cultural context: looking beyond your existing customer base to identify the broader barriers and opportunities for growth.

The primary distinction in using this data for strategic purposes rather than performance-based targeting, is that you are generally looking for broad trends and opportunities across aggregated segments of the population, rather than targeting at an individual user level.

Data providers use demographic, interests and purchase data from a range of partners to link behavioural data directly to sales; while social monitoring services provide insight into the conversations occurring around your business, brand and beyond.

Example 3rd Party Data questions

Ambition: How many people are currently interested in purchasing from my category?

Audience: What are the behavioural segments that are more likely to be heavy buyers?

Brand: What words or phrases are most commonly associated with my category across social platforms, forums and editorial sites?

Barrier: What are the negative comments most commonly associated with my brand?

Messaging: What video content was consumed by visitors to my brand website prior to their arrival?

Media: What content interests are most likely to indicate if someone will shortly be making a purchase?

Culture: What is generating talkability in popular culture at this very moment?

Applying 3rd Party data

There are a multitude of ways to apply 3rd Party data. From a social monitoring perspective, you can track conversations beyond your own social pages: analysing live content across platforms, forums and blogs to identify newsworthy topics with high social buzz: creating customised content in real-time to leverage a peak in consumer interest.

A social monitoring service can also help identify which of your products are generating high volumes of talkability and the key content themes. Feeding these insights into portfolio planning and creative development process can then maximise positive earned media from your paid advertising.

Finally, your DMP can bring together attitudes, interests and category shopping behaviour across a multitude of segments leading into and after a website visit or purchase. Looking at the ‘unrelated’ shopping basket items of customer prospects can also unlock insights into potential brand partnerships and cross-promotion opportunities, such as the (perhaps unsurprising!) increase in wine purchasing by expectant parents.

Case Study: Dove #speakbeautiful

Dove had been campaigning in the US with great success to change perceptions of beauty and women’s sense of self-worth for over a decade, meaning it was important to find ways to keep that message fresh and ensure the brand was continuing to deliver a positive impact.

With Dove’s research showing that social media had become the strongest influence on young women’s body confidence and self-esteem, in an environment where negative comments on personal appearance have sadly become the norm, there was an opportunity to initiate a positive behavior change in social media.

Using live social data, the agency found that Twitter was the most popular platform for beauty-related conversations, and that women were twice as likely to Tweet something negative about themselves, with 4 out of 5 negative beauty Tweets coming from women talking about themselves.

This was done by analysing every beauty-related conversation that mentioned appearance, body image or self-esteem in combination with either positive or negative terms. An automated algorithm allowed them to categorise these conversations for analysis and action, such as Tweets that were ‘self-oriented’ versus ‘oriented towards others’.

By looking at the seasonality and keyword associations with this content, they found that negative comments tended to peak around red carpet events, in particular the Oscars, with over 47% of Tweets in the negative. That’s more than three million negative Tweets around just one event.

From this, a powerful campaign idea was born, #speakbeautiful, aiming to change the way women speak about themselves on social media

Whenever someone tweeted something negative about their own appearance, the algorithm helped identify and classify those comments, allowing Dove to respond with a positive message either directly or by retweeting a positive message from one of their follower’s back to their account.

The results were impressive. Launching on Twitter three days prior to the Oscars - and supported by advertising during the show itself - the campaign generated over 5.9M Tweets and 801M Impressions, with 411,000 Tweets mentioning Dove by name.

Most importantly, however, there were 30% fewer negative tweets and 69% more positive tweets about self-beauty compared to the year prior, meaning the campaign had a fundamental impact on the way body image was discussed and perceived on social media, a truly remarkable achievement.

Behavioural Data Limitations

Like any data source, there are always going to be limitations, meaning a degree of pragmatism and restraint is required when applying insights to strategic decision making.

?      Behavioural data is typically unavailable for specific offline actions.

?      1st party data can be skewed by heavy digital users or regular customers.

?      Reporting methodologies for some 2nd and 3rd party data sources can be lacking in transparency.

?      Much of the 2nd party data is only available on request from the media owners.

?      There is often no ability to look at individual segments meaning insights can be generic to the total Internet population.

?      There isn’t yet a single customer view across all platforms to map the complete customer journey with confidence.

The Future of Data-led Planning

Despite all this, the fundamental output of the planner will remain unchanged: defining the challenge to be solved and using human intuition to convert observations into insights and implications.

What will change is the sheer volume of information we will have at our fingertips; the increasingly sophisticated optimisation algorithms aiding our decision-making process; and the frequency and speed at which we must adapt our thinking.

There are three broad stages in this evolution of data-led planning: aggregation; automation; and hybridisation.

1.    Aggregation

The first stage, which has been underway for many years, will see the continued expansion of integrated marketing dashboards designed to aggregate these disparate data sources, although with a view to making them more modular, customisable and accessible for broader strategic planning requirements.

2.    Automation

Over the coming years, algorithms will continue to proliferate and become ever more central to our planning process: identifying and extracting unusual or interesting developments in the data set; while the planner develops and deploys scenarios, responding in an adaptive manner as and when the need arises.

3.    Hybridisation

The future of advertising is often painted as a battle between man and machine, one in which humans will ultimately be supplanted by hyper-efficient algorithms, defeated one-by-one in the same way that Kasparov ultimately succumbed to Deep Blue.

And yet in freestyle chess today, it’s the combination of chess player and computerized chess engine – known jointly as centaurs – that have proved to be superior to machines.

In the same way, the planner and machine will work in a symbiotic union, with the technology identifying patterns and predicting the performance of various scenarios, while the planner sets and reframes the objectives, making abstract leaps of intuition to define strategy and execution.

In embracing the role of a ‘strategy centaur’, we’ll ultimately enhance our understanding of people’s relationships with brands, media and culture; sparking insights that will inspire even greater breakthroughs in strategy and creativity, ultimately bridging the worlds of short-term and long-term thinking.