Start with the end in mind: A three-step segmentation remedy

Segmentation has often faced accusations of failing to live up to expectations. All too often, these criticisms are warranted.

However, one area which often brings down segmentation initiatives is because the complexities and organizational discomforts of activation are postponed until the end of the process. If these realities are accounted for from the beginning - starting with the end in mind – there are specific actions we can take to ensure success.

It is not a matter of who owns and controls the best data full stop, but about identifying the right data for a specific purpose and how it can be connected to address the business question needing answered.

It is the combination of behavioral and survey data that creates a powerful starting point to build a strategy and activation across the business. This requires creating cross functional teams.

In some cases, this also means bringing those teams under one management structure, sharing ownership, and acknowledging the value of different data sources. But this also requires top management leading the process.

Without consistent support and management buy in, it is hard for those leading the functional teams to implement. We have heard from our own clients that there is a strong need for them to be supported by their superiors: it is all about leading by example.

It also requires training across the board, so the formerly siloed teams learn how the other team works, what they need, why and how to connect the dots.

These are just a few of the often-heard criticisms when it comes to segmentation, the latter of which has kick-started many a Request for Proposal (RFP). The biting criticism stems from having gone through a segmentation process where the outcome was felt to be disappointing: activation, business relevance, or effectiveness did not meet expectations. All too often, these are real and fair criticisms. In this article our aim is to advocate that we can develop any technically advanced segmentation, develop insightful and engaging reports, get all stakeholders on board to immerse into the segmentation, support in making the segmentation a part of any follow up project and develop a seamless process. We’ll end up with the same criticisms if in the end, the insights and segmentation are not implemented and integrated in the key data streams that feed the client organization. Make sure your marketing, sales, insight teams are connected with your data teams: have them work together, break the silos, make them part of the same strategy, and ensure that every study you do is connected to this internal network. We will demonstrate through a case study how we started by developing a cross functional team, and then used a structured process to ensure insights could be incorporated into the client’s database in order to leverage these strategic insights in the long run.

WHERE WE STAND TODAY

In a survey of 200 senior executives of large companies published by the Harvard Business Review, 59% of them reported having conducted a major segmentation in the last two years.1 Yet only 14% said they received real value from this exercise. Still, segmentation is on many brands’ radars to guide them in their growth strategy. Hence many papers, articles or POVs have been written on how to improve the effectiveness of segmentations. Often the solutions explored concern fixing mechanical steps in the segmentation process, meaning: the choice of variables, ensuring sufficient behavioral/real data, how to define the segments, how to prioritize them, etc. Another set of articles attacks the subject from a business application angle - observing failures to ensure the segmentation starts from the business objectives, engages stakeholders, drives focus and tracks impact. But what about setting the organization up for success with a segmentation? We can create a best-in-class segmentation and the perfect process around it, but things will likely fall flat if an organization is not designed or culturally primed for implementation of key learnings. We see this especially within a fairly new area of segmentation - precision marketing - which is all about activating segments to ensure personalized communication. Personal marketing is in essence sending consumers a communication about a topic that interests them personally, talking about it in a way that resonates with them personally, is sent in a format that they often use, through a channel that fits with them personally. A key challenge - where to find them! What media or channels should be used to ensure your message reaches them? Personal communication goes beyond developing offers and communication expressions tailored to specific needs. It also touches upon finding the right format, the right means and the right channel to connect with consumers. This paper will focus on the role of organizational dynamics in effectively addressing precision marketing challenges.

NOT MY PROBLEM

Whereas many segmentations are designed to set a broader framework for activities within the context of a brand or portfolio strategy, precision marketing drills down to putting the segmentation to the test: Have we chosen the right targets? Have we developed the right communication? Have we pushed the communication through the right channel? In short: did we do our job right? Precision marketing is all about increasing the ROI of your activities in general and messaging specifically. It is all about making sure you have a higher response from your consumers in a manner that contributes to your sales. But is it actually put into practice? Are you part of this process? There will be cases where the answer is: “No, it is not happening” … or at best “I don’t know. I hope so. I am not sure.” And how often do we postpone critical activation thinking to the end of the process? Sometimes we will justify our procrastination with this type of thinking: “The activation is not set up as part of this segmentation brief. It is a different team that is responsible for media buying and planning.” So what stops some companies from putting this process in place in the early stages of a segmentation initiative? What makes some companies more successful in this than others?

REAL ORGANIZATIONAL CHALLENGES

1. SILOS

While insights functions are usually well connected and integrated with the brand and/or product development teams, they are often separated from sales and media. Each team has its own data sources upon which strategy is built; therefore, they have different definitions of target consumers and how to activate against them. Insights teams and brand managers tend to base segment definitions on needs, attitudes, perceptions, and other survey data. Sales and media planners are often handed these definitions and asked to find these consumers in their databases (which typically store some combination of transactional data, socio-demographical data, digital data – but not the survey data).

2. DISCONNECTED PROCESSES

Various functions are sometimes completely unaware of activation activity elsewhere in the organization. Ignorance of one another’s processes can lead to misconceptions of how each partner could help in building a process front to end. Synergy between these teams and processes can be productive and create more value for the organization, but recent battles between big data and consumer data in some organizations have created a sense of distrust and protecting one’s territory. This only serves to extend an inefficient status quo. Things can get technical: it requires a lot of explaining on both sides how processes work with respective jargon. But the sooner various teams start working together, the better they will learn to anticipate each other’s needs to create a synergy instead of a battle.

3. PACE OF CHANGE & GEO-COMPLEXITY

The database world is ever-changing in its possibilities and regulations - what is available and what is possible. And there are differences in who owns and controls these data. One needs to wrap one’s head around the dynamic complexities of DMP, DSP, API, PI, cookies, and many more technical terms. For global companies, an additional complicating factor is added: differences by regions across the globe. If relying on digital data, the providers differ from Asia to North America, including the data these providers collect and store. To name one obvious example: Facebook collects data in many parts of the world, but not in China, which has its own data infrastructure. If you rely on transactional data, this will come from different partners as well: retail is organized differently even on a country by country level. The competitive offer is differently organized, meaning that data will be collected differently by country, if data are being collected at all.

SOLUTION: START WITH THE END IN MIND

Often, segmentation initiatives fail because we postpone the complexities and organizational discomforts of activation until the end of the process. But if we account for these realities early on - starting with the end in mind - there are specific actions we can take to ensure success. It is not a matter of who owns/controls the best data full stop, but about identifying the right data for a specific purpose and how it can be connected to address the business question. It is the combination of behavioral and survey data that creates a powerful starting point to build a strategy and activation across the business. This requires creating cross functional teams. In some cases, this also means bringing those teams under one management structure, sharing ownership and acknowledging the value of different data sources. But this also requires top management leading the process. Without consistent support, it is hard for those leading the functional teams to implement. We have heard from our own clients that there is a strong need for them to be supported by their superiors: it is all about leading by example. This also requires training across the board so the formerly siloed teams learn how the other team works, what they need, why and how to connect the dots.

CLARIFY THE DATA STRATEGY

We need to define upfront what data sources and data points are relevant to set strategy. Too many times we receive data files containing data gathered over time, without a purpose of how to use it – or with a lot of missing/dispersed data. Only collect data that you will use, that you can collect reliably and with consistency. Plan ahead how these data will be linked to other data, whether these are transactional data, external data, consumer insights or data you use for activation. All your data needs to be part of an integrated ecosystem. Train your teams as they will need to understand how these linkages work, and based on what. Knowledgeable teams will find it easier to integrate new data and connect the dots. Note: This will become even more important when we can’t rely on cookie-based data. Companies need to think about creating their own, first party data or create alliances with companies. Though this will require investments at the start, the efficiencies and synergies gained through collaborative teams and the faster turn-around of activities are only a couple of added values that will contribute to an overall better ROI.

INTEGRATED ANALYTIC PLANNING

It is not safe to assume data connections will be “figured out along the way”. The old way of running segmentations does not work anymore: bring the media teams on board, carefully examine their needs and the data they use for activation, and make sure these data points are integrated in the survey and even part of the analysis. These shared/common data points will be your keys to unlock the database. Your research supplier can ensure the segmentation is distinctive enough on these key data points and build algorithms more effectively. The same applies if you want to find your key targets in your CRM data. There is no magical formula we can develop at the back end if we have not properly planned the process upfront. Note: Data linkage can never be an afterthought. It needs to be planned carefully from the start.

GETTING IT RIGHT: A CASE STUDY

Microtargeting has proven to be an effective approach to add a human layer to the CRM database. Adding this human layer allows companies to not only understand customers, but to target each and every customer in a personally relevant way, with the right marketing action and the right tone of voice. In the case of VodafoneZiggo, the objective was to reduce churn (the number of people who switch to another provider). VodafoneZiggo collects many different data points, which gives them insight into the current behavior of customers and via an accurate model to predict customers at risk. However they felt there was still a knowledge gap to understand why people behave the way they do and how they could become more personally relevant and find a way to stop those at risk from churning. We worked with them to create targeted, behavioral science inspired intervention strategies and identified different types of customers in their database which were attracted to different approaches. This will better equipVodafoneZiggo to offer each customer the right solution, with the right tone of voice at the right time to increase loyalty. Thanks to microtargeting, we categorize people in the most granular way possible by blending several types of data – creating robust segments defined by attitudes, behaviors and activities. There were several key steps that proved critical on the path to success:

• First, we needed a cross-functional collaboration (led by Ipsos Strategy3) between the client’s marketing, insights, data analytics and data science teams to align on a variety of compelling intervention strategies to be tested. These were pragmatically ranked in terms of potential impact and achievability.
• With a clear analytic path, we then selected respondents from the company’s existing customer database - partnering with them to push our survey to a subset of their customers. We used MaxDiff to understand customer attitudes towards a telecom provider and their reaction to the various intervention strategies/stimuli.
• We found four distinct segments in the survey data set. Because we started with the end in mind - designing our analytic plan and survey in alignment with database realities - we were very successful at identifying and tagging these segments within the broader data set. We continue to work with VodafoneZiggo to refine the groups being exposed to various intervention strategies through continuous A/B testing.
From a technical standpoint, this is an example of integrating a company’s customer database with a survey to enable the programmatic scaling of segments to an entire database. But it was only made possible due to the collaborative planning between all the different teams from marketing to data science teams.

Most segmentation complaints center around implementation. Many companies complete compelling and intuitive market segmentations but then struggle with what to do next. Successful activation is often thwarted by some combination of silos, disjointed processes, geo- technical complexities – and lack of planning. But these challenges can be pre-empted if cross-functional teams get involved early and commit to a disciplined process with clear analytic planning/forethought. It is not only about connecting data but also connecting people within and outside your organization.