Embracing data-driven marketing is vital for business growth, but to achieve this organisations need to commit to constant experimentation and learning
20 December 2022
The evolutions under way in media today present unprecedented challenges for marketers. A proliferation of new channels is causing audiences to fragment. Along with this comes an influx of data from various platforms, possessing different levels of granularity, and adhering to different standards in the definitions and measures they apply.
There is potential for the cost and complexity of reaching relevant audiences in their preferred channels to balloon. But in this tough economic climate, marketing budgets and resources are unlikely to grow at the same pace.
There is also an opportunity in this fragmented media landscape to improve efficiency and overall effectiveness, however, by using a data-driven approach to budgeting, targeting and measuring campaigns. Indeed, by constantly learning and experimenting, the most successful brands can use data to drive growth even more effectively, connecting brand and performance marketing activities that are too often disjointed.
There are five key pillars to embedding this approach successfully.
Sponsored by Meta
By Albert Abello Lozano, head of automation, Treatwell
How to set up a marketing organisation for growth in today’s media landscape
Danish retailer Bolia achieved
Ultimately, brands that are achieving sustained growth understand digital channels can deliver both brand building and performance, as part of cross-channel strategies.
The route to full-funnel growth through cross-media campaigns is clear, if not necessarily simple. Silos – both of data and organisational structure – must be broken down. Teams must pull together and be judged on long-term effectiveness and profitability, as well as sales uplifts from performance campaigns. And they must readily share data and insights, collaborating on experimentation and optimisation, and validating the results.
This commitment to collaboration around data is not just a marketing challenge but also an organisational one – and one that is emphatically worth facing, as it can be the key to unlocking a new phase of marketing-led growth. ■
Learn more in the ‘Measurement 360’ whitepaper from Deloitte and Meta.
The role of digital channels is changing, with brands increasingly using them to achieve long-term as well as short-term marketing objectives. In this context, innovation in data-driven marketing across all channels will be a key part of unlocking future growth, and part of this will be establishing new tools and metrics at the heart of a marketing organisation.
Advertisers’ measurement framework should combine techniques for short-term tactical optimisation with longer-term strategic goals. Return on investment and return on ad spend are still key to direct-response optimisation, but more focus is warranted on long-term growth metrics that go beyond brand health - such as customer lifetime value (LTV), long-term ROI and price elasticity.
Brands have often defaulted in the past to focusing on the short-term effects of digital marketing, because conversions and sales are relatively simple to attribute to specific campaigns and the link to financial results is clear to the business. Marketing organisations may also have lacked the confidence and analytical skills to calculate LTV, use econometric techniques such as marketing mix modelling (MMM), or perform controlled experiments and lift studies.
However, the measurement tools available today put these approaches within easier reach (see section 3, below), which makes it more viable to take a long-term view with KPIs.
1. Set short-term and long-term KPIs
Now that third-party cookies are being phased out and users can opt out of tracking more easily, marketers are tasked with finding privacy-focused ways of turning data into customer insight.
If brands have the right data infrastructure and consumer consent in place, first-party data can be used to personalise experiences around where each person currently is in the customer journey. It is also crucial to calculating LTV, based on how often a customer makes a purchase within a typical purchase cycle, how much they spend each time, their projected spend over the whole brand relationship and the potential length of that relationship. Having this information allows brands to optimise campaigns for customer value against custom and lookalike audiences.
Application programming interface (API) technology has a big role to play, so that campaigns can be powered by an advertiser’s first-party data combined with that of ad platform partners in ways that respect consumer privacy. For example, within the Meta Conversions API, brands can use lift studies to test both upper-funnel reach campaigns and lower-funnel performance activity.
Danish furniture retailer Bolia implemented the Conversions API to understand the relationship between online advertising and in-store sales. Its media agency, Dentsu Denmark, helped structure and anonymise the brand's retail sales data and connect it to activity across its digital and physical channels.
Bolia could therefore measure the impact of Meta ads on both online and offline sales, and then optimise its campaigns based on the audience’s shopping preferences. The brand achieved a 4x increase in attributed sales, a 2.3x increase in ROI and a 23% increase in attributed website conversions as a result.
2. Invest in first-party data infrastructure
increase in attributed sales
increase in attributed website conversions
Just as new thought is going into campaign planning, brands are reappraising marketing measurement. This begins with gap analysis, to identify where they are currently failing to capture the long- and short-term impacts of their marketing. As a result of this, marketers are likely to realise no one measurement tool can serve all an advertiser’s needs today; rather a portfolio of them is needed.
To this end, Meta has worked with Deloitte on a new approach called Measurement 360, which recommends a holistic view of brand metrics and sales effects. For tactical optimisation, brands can test full-funnel campaign plans with lift studies, while marketing mix modelling allows them to measure against strategic goals, combining the short- and long-term impacts of marketing as well as factors such as price elasticity.
3. Set up your measurement for the evolving media landscape
What are my marketing goals, and do I have the right tools to best measure performance?
Define your goals, implement learnings from testing, define your measurement tools
Test and evaluate new marketing strategies
Which channels and media strategies best drive my marketing goals, and where should I invest?
Which data inputs and measurement solutions should I explore and add?
Marketers might perceive econometric techniques such as MMM to be the preserve of big brands with large teams, but econometrics is now a viable part of the measurement puzzle for any business. MMM can be automated and performed rapidly, saving time and resources. It is granular enough to meet different advertising needs. And it provides actionable insight that can support tactical decision-making on campaign optimisation as well as long-term planning and evaluation.
Attribution models can still be useful for measuring and comparing short-term effects of direct-response campaigns, but they can’t be relied on as the basis for budget-setting as they miss long-term impacts. The gold standard, therefore, is to triangulate with attribution, econometrics and controlled experiments, using the different methods to explain and validate one another.
The proviso is that business operations and organisational structures need to be set up for these new ways of working.
4. Champion a test-and-learn mindset
It is vital to evaluate different measurement methodologies, so brands can detect and analyse the incremental impacts of various campaign strategies. Testing new ways of targeting audiences throughout the funnel with different messages and ad formats can unlock new sources of growth.
Research shows companies that run 15 experiments in a year can achieve a 30% higher ad performance compared to those that do not. Those that have also previously run 15 experiments in the preceding year can expect to see a 45% lift.
by using Meta’s Conversions API
Companies that run 15 experiments
in a year can achieve
Those that also ran 15 experiments in the preceding year can achieve
higher ad performance
higher ad performance
Source: Meta And Duke University
This commitment to a learning agenda is essential for brands who want to take the best practice currently employed and then see how it can be improved through constant experimentation. Where better approaches are found, they need to be shared with the wider organisation and implemented, which means silos that prevent this need to be broken down.
And, of course, experiments do not always produce the desired results, so testing budgets and resources need to allow for this, with incentives for learning from failures and taking time to reach the right conclusions.
5. Build a team that can both interpret and present the data
Growth in today’s media landscape requires new kinds of analysis and new ways of communicating within organisations. These will be achieved partly through shifts in mindset and prioritisation (for example, more focus on long-term metrics) and partly through a skills evolution to include statistical modelling capabilities, for example.
Too often, brand and performance marketing are run by separate teams, competing for budget and not sharing insight. The customer’s experience, as well as business outcomes, will improve if data, marketing and product functions think holistically with the same goals.
Forward-thinking organisations are structuring their teams around a shared commitment to working and learning together, for both the short- and long-term future of their brand. Clearly, team members across all functions need to understand the business priorities, and be incentivised against targets that are designed to meet them. For team and business leaders, setting teams up with the right skills and motivations to work in this way will be key.
A template for data-driven growth
Source: Meta and DeloiTTE