Skip to content
Ideas Strategy2019-09-23

It's a bad time to be average: using data to differentiate your brand

Toby Roberts

Every company is chasing growth, but achieving growth is becoming harder — and more important — than ever before.

Download the full report: 'It's a bad time to be average'

According to McKinsey, only 8% of companies in any given 10-year period will grow significantly, but for those that do, the rewards are substantial. The top quintile of performers capture 90% of the profit in the marketplace and establish the margins that allow for long-term marketing investment.[1] For the rest, strategies become “incremental improvements that leave companies playing along with the rest of their industries,”[2] leaving lots of brands stuck in the middle, with low growth and lower profitability. This is why it’s a bad time for any brand to be average.

In marketing and communications, we see the same. The growth and availability of audience data and media technology is leading to the unthinking pursuit of the same strategies and tactics regardless of the company or industry. This approach is pushing brands toward average at a time when consumers are increasingly seeking differentiation. Marketers are looking at the same data, in the same way, and reaching the same conclusions. As a result, everyone is chasing the same people, with the same messages in the same channels — focusing budgets on extracting efficiency at the bottom of the funnel.

In a time of immense change, it is important to remember that two fundamental truths in marketing remain the same:

  1. Brands help companies grow. Recent research from the National Bureau of Economic Research shows that over a third of the value of high-growth companies comes from brand value. Almost as much as they get from their IP.[3]

  2. Brand differentiation is key to predicting future brand growth.

These truths remain unchanged. Brands are part of society and culture and therefore should look to harness (and sometimes change), cultural norms and perceptions. They are “mass” collective constructions — what a consumer thinks about a brand is as much conditioned by what other people think about it as it is by what that brand tells them. 

However, as there are now fewer mass media platforms on which to stand and get everyone’s attention, this has become harder, and more expensive. Instead we have embraced new opportunities for targeting and personalization which are opening up exciting possibilities for communication that would have been unthinkable just a few short years ago.

These approaches can be very successful, but there is an increasing view within the industry that the pursuit of these narrowly targeted opportunities has come at the expense of bolder, differentiated strategies that build long term value for brands and businesses. This is supported by research from the Institute of Practitioners in Advertising in the U.K., which shows the need for a better balance between the focus on short term efficiency and long term communication effectiveness.[4]

Navigating a “de-averaging” world

It shouldn’t — and doesn’t have to — be this way. The explosion of media choice and changes in retail and purchasing behavior give marketers more opportunity than ever before to do different, interesting and innovative things to differentiate and grow brands. Not only that, these changes in media consumption and purchase behavior are creating much more nuanced and granular data that — when combined with more “traditional” datasets — allow for greater experimentation and diversity in our approaches to building brands for the long term while still harvesting demand in the short term.

By paying attention to a broader range of data signals across the whole marketing landscape, we can harness the new capabilities afforded to us by data and technology and combine this with what we have always known about brands.

01. Building models to make strategic choices more tangible

The challenge

Brand differentiation must start with an ambitious marketing strategy. Too commonly, however, marketing strategy is constrained by the need for communication investment to focus on short-term returns with advertising that either pays off immediately by converting in-market audiences, or later by building awareness, consideration or affinity among a broader out-of-market audience. 

This approach limits the ability for marketers to create differentiation at a time in which media is increasingly the channel in which people experience and decide to buy brands. Deprioritizing marketing’s role in creating long-term strategic differentiation not only produces incongruity between our actions and the actions we believe are necessary to be competitive in the future, but means gradual marketplace changes often accrete unwatched until they become urgent disruptive threats.

Regardless of the category (books, retail apparel, consumer electronics) recent closures and bankruptcies point to the ease with which companies forget, or de-prioritize the role of marketing on future growth. The challenge is to make the long-term view tangible today, enabling the exploration and potential impact of bold strategic moves. 

The solution

It’s critical to stop viewing advertising in isolation and start thinking about it in the context of everything else the brand is doing and other forces affecting the category. Doing this allows marketers to evaluate strategic trade-offs shaping the future.

The answer is to use data to build a “world” in which we can test different strategic hypotheses and explore different scenarios — quantifying the effects of different potential marketing interventions on both the long and short term. This approach makes the impact of achieving — or not achieving — long-term growth tangible and real by placing a value on brand differentiation and enabling prioritization alongside short-term goals.

This sounds incredibly complex and difficult but in truth, most organizations and their agencies have sufficient data, experience and knowledge to be able to do this in a structured way. It’s simply a case of adopting new approaches.

Example: Agent Based Modeling and Marketing Simulations

Despite the explosion of data and new analytical methods over the last few years, the tools to evaluate strategy have remained stubbornly the same — marketing mix modeling, segmentation, conjoint analysis and spreadsheets. It’s a surprising lack of innovation given the dollars at stake.

In many ways, it is no wonder that these legacy measurement tools don’t adequately reflect the long-term impact of marketing strategy. They don’t capture why it is important. It is not just how long the effects of advertising endure that matters but the value it adds to all aspects of the business and its defense against competitive disruption. To model this complexity requires a different approach.

Agent Based Models create a simulation of the marketplace — a “Sim City” of marketing. They allow us to “test” strategies in a virtual view of the market. And by making the consequences of our strategies more tangible, it is easier to be bolder and more committed across multiple stakeholders.

Test scenario examples:

  • Impact of new brand positioning on sales

  • Changing prioritization of different segments due to competitive disruption

  • Impact of new e-commerce models on advertising budgets

  • Optimal mix of broadcast and addressable media

Unlike existing techniques, marketing simulations emphasize the importance of consumer diversity. They don’t take a complicated consumer journey and assume it applies to everyone. They don’t try to force millions of people into five segments. And they don’t make the assumption that consumers behave rationally. In a simulation, consumers can think both “fast and slow.”

Market simulations take full advantage of the explosion in granular, individual level data to create accurate views of the market. They are complementary to nearly all measurement methods, aggregating intelligence from every angle. For example, Essence helped a client reposition its brand portfolio using a simulation built with respondent-level purchasing data. By linking brand positioning and sales, we were able to evaluate different targeting strategies and guide channel splits and flighting. The result was a bolder strategy that reflected the client’s unique situation and the competitive threat. We have built similar models in telecoms, automotive, retail, durables, CPG, finance and healthcare.

02. Using tools and technology to plan away from the average

The challenge

Media planning tools, most of which offer some combination of budgeting, modeling and optimization, are necessarily built on past data. They are predictions based on an average of past performance. The problem is that many great ideas and strategies get lost along the way through trial-by-analytics. Ideas that would have led to something genuinely different get dragged back toward the average precisely because the benchmarks that are meant to help validate them cannot accommodate them. This means that tactical plans can end up looking very different from the intended strategy and very similar to the previous tactical plan.

The solution

Use the historical data and analytics to rapidly develop a “base plan” early in the process. This is the “average” of everything that’s gone before in the category and represents the best aggregate answer. This is then your foundation — through the lens of the brand and its specific challenges, strengths and weaknesses — build on this plan with specific contexts, moments, occasions and any other broader opportunities that serve to both enhance the plan’s performance, as well as make it specifically ownable for a brand. 

Off the back of this principle, Essence is building both general and client-specific tools that ingest historical data and then create the most efficient base plan for a given brief, based on past campaign performance. These tools can quickly optimize a “generic” plan based on what has happened in the past — everything from budget setting and allocation to keyword identification, channel level reach optimization, and more tactical publisher- level allocations  for direct response planning. We then challenge ourselves to reimagine this plan based on the specifics of the brief in front of us — to identify the vectors of differentiation that will accelerate growth, and build ideas around those. 

03. Separating signal from noise when defining audiences

The challenge

The growth in granular audience data is opening up possibilities in communication that were unimaginable just a few years ago. But more data doesn’t necessarily lead to greater clarity — there is a danger of becoming so obsessed with what is known about the audience, and the assumption of what they want to hear, that marketers forget to prioritize what the brand  actually wants to say. This is further exacerbated by the common myth that brands have their own unique, distinct target audiences. The growth in audience data is fueling this delusion. With very few exceptions, it is false. Categories have target audiences, but brands — by and large — do not.[5]

This leads to a concentration of category spend against either smaller audiences, as everyone chases the few “hand raisers” in their category, or against the broader contexts and “passion points” in which the audience shows a general interest.

The solution 

Ultimately, a marketer’s role in growing the brand is to make people take an interest in what that brand has to say. Not to relentlessly target people who are already interested, or to talk to them while they’re doing something else they happen to be interested in.

The first step to getting this right is to treat targeting and segmentation as different but related tasks. Defining a target audience is based on the objectives being set for a campaign or other activity, the life cycle of the brand, and the competitive context of the category. It should start broad and then be populated with data, not vice versa. It’s unnecessary to narrow down an audience at the start of the process, and it can hamper growth ambitions if the audience is whittled down too much.

Once this group has been defined, then think about how to segment it into smaller groups, remembering that a segmentation is only meaningful in the sense that it should clearly indicate different communications approaches for different groups. This means that, usually, it should be based on category and brand needs and behaviors, not differences in demographics, affinities or interests unconnected with the category in which the brand is operating. 

This isn’t a new approach, but there’s a danger of it being overlooked as marketers are consumed with brand-level user data. Also, it can often be hard to achieve — the data may not exist, or be difficult to acquire, for example — but if the desire is building differentiated brands, marketers have to demonstrate relevance to audiences in the long term, as well as harvest demand in the short term. That means focusing on what the brand wants to say, and how they want to say it, rather than imagining what people want to hear.

04. Applying data to make every interaction relevant (not necessarily personalized)

The challenge

Going hand in hand with these greater targeting capabilities is the increasing use of data to personalize advertising messages to make them more relevant to consumers. There is no doubt that, when done well, such approaches are highly impactful and can significantly boost campaign performance. 

As discussed above, audiences should be segmented based on the objective for that particular group, while cognizant of their attitudes and behaviors toward the brand and the category it is in. These critical segments can then be overlaid with richer data in order to provide more texture and nuance for creative development.

The problem arises when marketers jump straight to the second step and neglect the first. If there is no understanding of the specific brand objective against the audiences being considered, then any personalization of the communication will be arbitrary and irrelevant.

The solution

The key is to identify and prioritize those vectors that are most predictive of category or brand interest and build the brand’s messaging accordingly. For high interest, more complex purchases, the starting point for personalization should be an understanding of where an individual is in the funnel. This defines at a broad strategic level what the takeout of the message should be. Individual creative treatments can then be developed based on different audience segments (e.g., identifying the best product point to emphasize at the consideration stage and which price point to communicate at the buying stage).

However, this approach starts to break down in categories where there are fewer meaningful audience signals that predict receptivity to communication. In these instances, it is important to  remember that demonstrating an understanding of the audience is only one way of establishing relevance, and rather than simply using whatever data is available on different audience segments, marketers should explore different predictive data sets around which it is possible to tailor communication.

For example, brands, products and services are situationally relevant. To use an extreme example, creating a series of customized messages targeting homogenous segments for an ice cream brand is likely to be far less effective than targeting a broad range of consumers with a single message when it’s hot. Additionally, repeated studies have also shown that contextual relevance (i.e., making advertising more tailored to the context in which it appears) boosts recall and ROI.[6]

The best solution is to utilize as many of these different techniques as the plan will allow, without compromising the overall reach of the activity. This will allow the refining of activity over time, as well as the identification of untapped pockets of potential growth.

Example: Using custom algorithms to drive higher-value Google Store sales

Google and Essence use machine learning and Google’s own programmatic platforms to improve the performance and relevance of their advertising. Out of the box, tools like Display & Video 360 Custom Bidding (in beta) enable automated bidding strategies for a number of category generic KPIs. By leveraging first party data, and the results of a decade of effectiveness experiments, our data science team was able to upload a customizable scoring model to optimize bids in real-time against our brand-specific KPIs instead. Powered by machine learning, this approach allowed us to automatically score and prioritize our bids based on expected value at the impression level. The performance improvement was significant — driving stronger brand lift at a lower cost and demonstrating the power of AI when applied against measurable — and even granular — business goals. 

Summary: Creating growth through data — and differentiation

The compounding benefits of growth and investment in growth, and the connection between brand differentiation and brand growth make it a bad — and even dangerous — time to be average. The focus on short-term and unthinking audience planning is constraining brands and businesses from realizing the true value of communication to build differentiated brands over the long term.

Overcoming these challenges means shifting to a “de-averaging” mindset and behaviors. Marketers should continue to embrace the new possibilities created by data, technology and digital media, but reconfigure our approach around the unchanging “ truths” of marketing — the value of strong and differentiated brands to capture both short-term sales and build long-term value. This means focusing on a broader range of data signals that are outcome based and predictive of the future. These will continue to include the signals of audience relevance that have been so important in driving media efficiencies, but enhanced with more brand specific signals of growth and differentiation that will ensure longer term success.


1. Bradley, Hirt & Smit, 'Strategy Beyond the Hockey Stick,' McKinsey & Co, May 2018

2. Ibid.

3. Belo et al, 'Decomposing Firm Value,' NBER, July 2019

4. Binet & Field, 'Media in Focus: Marketing Effectiveness in the Digital Era,' IPA, October 2017

5. Byron Sharp, How Brands Grow, OUP, 2010

6. Kwon, et al, 'Impact of Media Context on Advertising Memory: A Meta-Analysis of Advertising Effectiveness,' Journal of Advertising Research, April 2018