Build the fundamentals including the team, processes, and analytics, identify the opportunities through self-declared, deductive, and inductive approaches, prioritize the efforts as there will be limited bandwidth in dealing with all the projects you may have in mind
Personalization is a popular topic in marketing and knowing the techniques for creating a professional personalization program will make a huge difference in marketing results, as sometimes it works and sometimes it just doesn’t. The goal of personalized marketing is to help improve conversion rates through customer segmentation rather than dealing with averages.
Furthermore, personalized marketing helps build and enhance efforts around customer differences. For example, if a segment of your website’s visitors responds better to a marketing message than another, are there better, more customized messages that would work for each of them, and therefore, should we split our marketing message and overall efforts into two segments or more.
This is what is called the Marketing Customization Spectrum where at one end, we have mass marketing, where everyone sees the same message which is best used for general awareness, brand building, and positioning while at the other end of the spectrum is one-to-one messaging, where each visitor of a website gets an individualized message across their touchpoints — which is the dream of one-to-one personalization that direct marketers. And in between these two extremes, is some variation of customer segmentation where groups or batches of customers receive their message and communication strategy.
Use personalization to create relevance
And the answer is it depends and you need to test it. No matter what grand scheme you come up with, you will need to test it. The best way to implement personalization is not to look for a bunch of tips and tricks and just implement them. Personalization ideas are only hypotheses until they’re tested. Different customers, different segments, will react differently to your campaign. So you will need to set metrics based on your goals, test your scheme, uncover customer reactions, and pivot or finetune your tactics.
The purpose of personalization is to make our prospect’s experience from their perspective more personal, more relevant, and more meaningful. In general, six factors improve a customer’s conversion:
- Your value proposition
- The relevance of your communication
- The clarity of your communication
- Lower levels of external distraction such as your competition
- Reduced customer anxiety through an optimized customer experience
- The degree of urgency the customer feels to satisfy their need
LIFT Model: Most Popular Conversion Optimization Framework
In 2009, Chris Goward introduced the LIFT Model®, a framework for analyzing web and mobile experiences and developing…
Personalized marketing helps create relevancy in your communication. Relevance to their needs, search keywords, the seasonality that they’re in, their geography, purchase intent, and so on. In short, personalization is used specifically to help solve the relevance problem.
Drivers of personalization failure
There are several drivers of failure to personalized marketing efforts and some of them include:
- Lack of an overarching marketing strategy. Personalized marketing is not a strategy but a tactic. Ad hoc implementation of off-the-shelf ideas using digital technologies will not bear results unless you have a strategy. You should first be asking, “What do I need to fix in my customer’s experiences?” before going out to communicate the improvement, and then when you do, you need to tailor your communication to be more personal.
- Lack of an A/B testing process will fail your personalized marketing efforts. Personalized marketing ideas are only hypotheses that need to be tested in action and therefore A/B testing should be the layer over top of all marketing hypotheses.
- Lack of insightful and meaningful data. To understand your customers, we need to hear from them directly and therefore it helps to have a structured process for talking to them and collecting true customer feedback.
- Lack of an experienced team without central coordination. The central planning hub for personalized marketing efforts should also be prioritizing where to spend capital to get the maximum return.
- Lack of resources to monitor all the variables that come from personalization activities. Varying messages across different customer segments or experiences will incur the cost of having to maintain the cost of updating all of those messages every time we publish a new release, a new product, a new season, or anything else that impacts our integrated marketing communication plan. Furthermore, we have to make sure that any personalization tactic is worth more than its costs and that can be predicted if/when we have the data.
Preparing the fundamentals
Before I dig any deeper into how to do personalization, you need to first determine whether personalization is the right tactic for your company right now. Let’s look at three basic questions that will help you figure out your personalization maturity and fit.
- Having the right team in place. Personalization efforts require a team that can manage the marketing strategy, customer segments and messaging, and the required tools. Personalization efforts add costs and require some specialized skills.
- Having processes for validation of personalization ideas. Personalization is a hypothesis until it’s been tested. These hypotheses are all assumptions about your customer segments until they’ve been validated. Successful personalization strategies are about having a rigorous process that enables gathering insights about customers and then validating those insights.
- Having insightful data and analytics in place. Developing a personalization strategy requires an audience-centered approach. Successful personalization programs will need to derive meaningful insights from their Google Analytics or any other analytics tool to solve customers’ conversion barriers. It’s best to understand the basic segments first and then layer on a personalization practice.
Once these elements are in place, a great team with processes and tools will create a virtuous cycle that will lead to even greater customer insights leading to improved expertise, knowledge, and effectiveness.
Building the team
One of the reasons personalization programs fail is a lack of resources, a lack of an expert team who can manage it.
Great personalization teams are great optimizers and experimenters. They need training and experience in a variety of subjects including personalization, A/B testing, behavioral design, cognitive biases, persuasive copyrighting, web technology and analytics, statistics, design of experiments, project management, and, marketing.
In addition to the required skillsets, they need to be result-oriented, consistently delivering business results, and therefore the team’s incentive structure should be base on this attribute and the quality of outputs. Furthermore, this type of meritocracy ensures a consistent team of players based on the objectives the team should have. To deliver on this, the team needs to stay sharp, always be doing research, keeping up to date on consumer psychology and behavioral economics and persuasion research, tapping into identifying whether campaigns apply to different situations and segments.
Furthermore, the in-house team can be supplemented with a specialist agency to speed up results delivery and the learning process. At times, you can choose to outsource some projects temporarily or permanently, to increase internal focus.
The process: focus on growth
The processes within a personalization marketing team should focus on growth marketing. Personalization optimization is about understanding more about customers and making the experience of the value proposition more relevant every day. And to succeed, it needs to explore and validate hypotheses. Exploration and validation are separate processes that are best not to happen simultaneously.
In the exploration phase, the goal is to analyze customer information such as those of web analytics, that reveal patterns of how customers experience the value proposition. These patterns need to lead to persuasion principles and cognitive biases, that enable a communication strategy considering the business context of the organization, which takes into account the brand history and product mix. Furthermore, as the number of insights and initiatives scale, the organization needs to build an archive or library of efforts to document the history of learning from previous experiments and to potentially replicate or optimize for future explorations. The exploration phase also needs user research, where the team gathers customer feedback and develops customer experience models. Exploration efforts will get the team to step in the shoes of the customer and understand where their conversion barriers may be across the six factors mentioned previously as the drivers of customer conversion.
Exploration generates insights, which need to be validated. The first step is to prioritize exploration hypotheses and allocate resources to the highest impact projects. Then, these high opportunity projects need to be tested in controlled testing environments. To maximize output and return on experiments, it is best to test variations of teste through A/B testing. And then the experiement goes live with real audiences which may require UI and UX design and technical development depending on the scope of the experiment. These live experiments need to be monitored for customer behavior and conversion rates. Once the live experiments have run for long enough to achieve the required statistical significance, results will need to be assessed in-depth and consequent decisions to continue, abandon, or pivot will be made.
And this loop of exploration and validation will keep on going forever, because of market dynamics such as your competitor’s actions or reactions.
The Complete Growth Marketing Playbook
An essential collection of mindsets, tools, and strategies in nine chapters
The Infinity Experimentation Process℠ | Widerfunnel
The most advanced organizations view optimization as an ongoing strategy. It is not a one-time project and it doesn't…
Investing in the tools
Technology and tools are essential to enabling a personalization strategy and initially, teams will need to integrate and share their data. There are two types of data to consider:
- Exploratory data, including research studies, user test findings, general web analytics insights, etc. This data can be qualitative or quantitative. This data will help build an understanding of existing personas, customers, and shifts across the industry.
- Dynamic data, which is live data about customers that help the personalization system create individualized experiences. Ideally, this will have a single view of the customers with accurate measurements so that the team can build an accurate view of how well your personalization hypotheses are working out.
To make life easier, you will need to set up an insights platform that takes large amounts of data and turns it into actionable insights for real-world experiments. There a variety of platforms available for this purpose including Qubit, Dynamic Yield, Evergage, Optimizely, and Sitecore. Therefore it is important to make a thorough assessment of how the tool will help you attain your goals before making a purchase.
Identifying opportunities for impact
Start with current customers
The first place to look for personalization insights is within the current organizational knowledge or current customer base which includes what your business knows, or thinks it knows, about your shopper. This is the starting point. The goal will be to identify who the customers are and what they’re interested in which can include the existing personas, influencers, and decision-makers, varying product interests, different stages in the buyer journey, seasonality habits, etc. The output in this stage can be a few customer segments with some applicable personalization hypotheses — these are the opportunities when the team can personalize the customers’ experiences right away; immediate impact. And the rest becomes testing those hypotheses and whether they will convert and stick.
Here, the team will need to have the experience and tools in place to uncover insights that can come from assessing search terms that show product interest or buyer journey stage, the ad campaigns they’re coming in on, geography, behavioral patterns on the website, etc.
Finding new insights
After exhausting current user information, the next layer is to move into new research insights where we can find new data or run different focused experiments to generate new personalization insights. Once again, we will need to be testing and validating which insights provide a return, and after having proven that the personalization approach is working we can move towards automating the procedure. Through automation, we embed discovered insights into the organization through a personalization campaign management platform. A lot of mistakes will happen if we move directly to automation before validation.
There are three ways we can identify new opportunities:
- Deductive-identified (top-down)
- Inductive-identified (bottom-up)
Self-selected personalization is where through product design or direct reach outs we provide customers with opportunities to identify themselves across our marketing experience and at which stage they are at so that in effect we can tailor and customize our messages and offers to be more relevant, and useful, and helpful for them. There are two factors to consider here:
- How do we best segment customers to find the most appropriate, relevant, and meaningful customer personas? There could be any number of ways to segment and personalize.
- Within each of these segmented personas, what are the best offers, and calls to action, or content for each of the segments?
Both of these areas should be tested and validated to ensure we’re taking the right approach.
Deductive methods are top-down approaches to testing personalization hypotheses that come from testing whether known general personalization or persuasion approaches can apply to other situations. For this to work, you will need to build and maintain a personalization campaign archive or library to quickly explore, test, and validate the approaches when circumstances come up. Make sure to isolate a control group to test against any personalization methods, especially if it’s not clear how the business rules will work yet in this case. Furthermore, use deductive reasoning to make a case for why the method should work in this particular situation before spending time testing something.
Through inductive investigation, we aim to find patterns by observing individual behaviors within available data and relating them to potential theories that could provide a robust understanding of the overall customers. Inductive analysis is a bottom-up approach to uncovering insights.
We need to consider that different messages can work for different periods in different seasonalities, which is segmenting by seasonality. e-Commerce platforms have specific campaigns and webpage designs for holiday seasons, tailoring their messages to the seasonal occasion with the call to action (i.e. to buy) shifting higher up on the page to give better access as customers have urgency to decide and make purchases, while during a low urgency season, e-retailers tend to emphasize the story and its assortments more.
Running personalization programs can have very high maintenance costs because once executed, you have to maintain all of those personalized messages and campaigns and when we increase the number of our segments and campaigns, the complexity can grow very quickly. That’s why in addition to automation, we need to prioritize efforts.
To find the best return on efforts it’s best to assess three levers:
- The potential: which ranks how much improvement in the user’s experience is possible.
- The importance: looks at how much impact to the organization this idea could represent.
- The ease: takes into consideration how much difficulty we could run into in implementing the opportunity.
How Can Marketers Prioritize A/B Tests? An Overview of PIE/PXL Methodologies | Something Digital
Not too long ago, we heard from my colleague, Lindsay Pugh, about the importance of A/B tests every change you make to…
Winning with growth and gathering further insights
Further to prioritization, we need to test personalization efforts because as much as we may believe to understand customers and the effectiveness of how we are communicating, customers tend to perceive our voice differently. The messages we send to customers are filtered through the customers’ perceptions and the only way to understand which messages work for each of your customers or segments, is to talk to them and test how they respond. The secret to a great optimized personalization program is to understand that there are two goals to every experimentation:
- Growth — i.e. getting winning results
- Insights — i.e. getting data about the customers' mindset to build new layers of experimentation
With the mindset to grow and further test new experiments, we can build a marketing loop that can only grow and strengthen our personalization efforts.
Personalized marketing is about creating a professional personalization program that helps improve conversion rates through customer segmentation rather than dealing with averages. It helps build and enhance efforts around customer differences as it creates relevance in communication. To be successful in personalized marketing we need to:
- Build the fundamentals including the team, processes, and analytics
- Identify the opportunities through self-declared, deductive, and inductive approaches
- Prioritize the efforts as there will be limited bandwidth in dealing with all the projects you may have in mind