What is personalization for marketing?
- What is web personalization?
- Manual personalization or predictive personalization (AI)?
- Which method should you choose?
- AI-driven personalization, predictive algorithms: how do they work?
- A/B testing and personalization: 2 complementary practices
- The benefits of personalization
- What sectors benefit most from personalization?
For organizations with established A/B testing programs, personalization can maximize how you use customer data to create standout online experiences and products.
Consumers now expect platforms, websites and apps to be frictionless, and personalization uses customer profiles and behaviour to tailor the customer experience to different users.
Personalized online experiences should reflect users needs, behavior and desires, thereby creating more engaged customers and conversions.
Imagine you’re an e-commerce website. You already use A/B testing and experimentation to hone your messaging, design and website structure. Done correctly, this should have already grown your revenue compared to deciding things on gut feeling alone.
However, you believe you can segment your digital experience further using customer data.
- Maybe you have separate catalogues for men and women
- You know that offering seasonal products to users near the equator is costing you sales
- Perhaps you want to display your Valentine’s Day sale to people more likely to be married
- To drive retention, you can create special messaging and prices focused on existing customers
Personalization lets you address visitors’ expectations and needs at each step of the customer journey and increase brands’ conversion rates and revenues.
This is why personalization has become central to a well-thought-out, customer-centric digital marketing strategy. A high number of businesses have already grasped its significance: 87% of marketing leaders said they included personalization in their strategy.
What is web personalization?
Personalization is a marketing strategy that involves contextualizing the browsing experience, messages and offers on a website, based on visitor characteristics.
It is based on analyzing visitor behavior and using this data to create tailored online experiences that improve conversion.
A personalization strategy has four stages:
- Identifying and segmenting your target
- Triggering a personalized action in real-time
- Managing marketing pressure
- Analyzing your performance using clearly identified goals
- The 4 stages of successful personalization
- Build the most effective personalization strategy: A 4-step roadmap
- Tactical Personalization is dead. Long live strategic personalization
- CRO Best Practices: Personalization part 1
Manual personalization or predictive personalization (AI)?
There are two methods when it comes to personalization. The method you use will depend on the strategy you adopt.
With manual personalization, the marketer manually segments their audience according to criteria they choose. These can be based on the results of previous data analyses or on simple observations.Once these segments are created, the marketer sets an action that is triggered for visitors within that segment.
Manual personalization is a good option to use if you have a specific idea of the customer segment you wish to reach and the criteria that define it:
- “I want to reward loyal customers who have spent more than $200 over the past six months, by inviting them to an event.”
- “I want to push a welcome offer to my new visitors.”
- "I want to reduce cart abandonment by adding an extra for those who have over $100 of products in the cart."
Personalization based on predictive algorithms
This automated approach involves relying on an artificial intelligence (AI)-based tool. The algorithm creates a real-time propensity score for each visitor according to their probability of converting to a pre-set goal.We call this process predictive targeting.
The machine learning algorithm analyzes the data of all the visitors to a website and establishes correlations between them. By doing so, it develops a capacity for prediction, enabling it to measure the conversion intention of new visitors.
Kameleoon needs 15 seconds to accurately determine the conversion probability of a new visitor based on their behavior and the context of their visit.
This conversion rate intention, displayed as the Kameleoon Conversion Score or KCS™, is what is used to target visitors. In real-time, we can identify the hot, undecided or cold visitors and trigger differentiated actions at an individual level.
If the goal you wish to achieve is difficult or even impossible to define by setting criteria in advance, because these criteria are vague, ambiguous or shifting, you can use predictive targeting to personalize.
Which method should you choose?
- When you can identify that a visitor belongs to a pre-determined segment.
- 1-to-few approach: your actions are addressed to a specific visitor segment.
- Targeting is static: it depends on rules set at the beginning and only evolves if someone changes these.
- The criteria that define the target can be clearly identified and modelled.
- When you cannot define the segmentation criteria in advance and you want to measure the conversion intention of each of your visitors.
- 1-to-1 approach: your actions are personalized based on how each visitor’s probability of conversion evolves.
- Targeting is dynamic: prediction evolves over time as the behavior of visitors on the website evolves.
- The criteria that define the target are difficult or impossible to model manually or depend on numerous or very disparate data.
- Why AI makes the human touch even more important in personalization
- AI Personalization: Adopt a predictive approach to create unique experiences and maximize conversions
- Dynamic Traffic Allocation vs Predictive Algorithms: Making Personalization Work
AI-driven personalization, predictive algorithms: how do they work?
AI-driven website personalization uses algorithms to target visitors and trigger experiments that are adapted to the visitor’s requirements in real-time.
Collecting visitor data and learning
Visitor behavior data is collected and analyzed by the algorithm and stored in the visitor’s browser. Only anonymized data is collected.
This hot data can be complemented with cold data from your CRM or DMP.
The AI algorithms learn from the behavior of visitors on your website by making correlations with already established profiles. They improve their prediction capability over time.
The algorithms’ predictions are defined by the marketer, depending on their goal. For example, this could be identifying a visitor’s conversion probability or interest.
The predictive algorithms used in personalization are based on machine learning. They are different from dynamic traffic allocation systems which use a statistical approach. Machine learning algorithms are based on different types of learning: supervised, unsupervised and reinforcement.
At Kameleoon, we use all three algorithm types:
- Supervised algorithms to identify behavioral trends among your website visitors
- Unsupervised algorithms to create visitor segments based on purchasing behavior or to create models. For example, a model might be that visitors that bought a certain product are very likely to behave in a particular way.
- Reinforcement algorithms to calculate visitors’ purchasing probability/propensity score based on their on-site behavior.
Triggering personalization experiences
Once the algorithm is able to identify visitors’ needs, it can trigger, in real-time, actions defined in advance.
When a visitor lands on the website, the algorithm analyzes their behavior, calculates a conversion probability and uses this to display the most suitable experience.
- Kameleoon's partners
- AI Personalization: How it really works
- An architecture designed for real-time and predictive actions
- A powerful AI engine
- Why AI personalization needs to open the black box to increase adoption
A/B testing and personalization: 2 complementary practices
In general, a user experience optimization strategy begins with creating general usability improvements for all visitors thanks to A/B testing.
Although necessary, these only cover part of the potential for conversion optimization that exists on each website.
Personalization covers a much wider spectrum, as it can be used to optimize the entire customer journey by creating customized experiences for each visitor or visitor segment (micro targeting), unlike A/B testing, which consists in making a usability trade-off that benefits all of a website’s visitors (macro targeting).
Let’s look at an example:
- if we choose the best possible variation for the three segments (variation 1 here) we see an average +3% increase in conversions (being an average of 2%, 4% and 3%)
- if we choose the best possible variation for each segment independently we see an average +4% increase in conversions (being a average of 3%, 4% and 5%)
The two practices should therefore be integrated to increase website conversions and offer visitors relevant and seamless experiences.
This means that you should work with a comprehensive solution that can handle both concepts, in order to deliver a complete approach to optimization.
- Is web personalization an extension of A/B testing?
- A/B testing, segmentation and AI personalization: 3 complementary approaches to conversion challenges
- 20 ideas for personalization and A/B Tests based on real case studies
The benefits of personalization
The better you know your customers, the more you can segment your audience and thus increase the value of each visitor by offering them increasingly personalized experiences. Personalization is a crucial way to address the strategic challenges facing all marketing managers.
BOOST YOUR CONVERSIONS
Personalization enables you to meet each visitor’s expectations, trigger conversions and also increase performance (by improving the average customer spend) or accelerating conversion speed thanks to delivering the right messages at the right time.
By offering targeted content and products, the browsing experience becomes more relevant and increases visitor engagement on your website. It can also make your content marketing more data-driven and intelligent.
Personalization enables you to identify the visitors who are about to leave your website, allowing you to trigger the personalized action that will keep them there at just the right moment.
GENERATE QUALIFIED LEADS
By contextualizing each visitor’s experience based on where they are along the conversion funnel, you can make it a seamless experience for them to provide their contact details, make an appointment or take an action that will engage them with your brand. This can drive customer acquisition.
Personalization is not just for the landing page, but end-to-end customer relationship management. Personalization enables you to build a trust-based relationship with your visitors and thus to retain them for the long term.
BOOST PROMOTIONAL CAMPAIGN ROI
With personalization, the impact on ROI is immediate. By correctly targeting offers and messages, it is possible to reduce acquisition costs by 50% while income increases by 5 to 10%. Marketing campaign spend is optimized while you obtain better results for the actions you take—for example, discount coupons are solely targeted to undecided visitors rather than being offered to those who would convert anyway.
With AI-driven personalization, the benefits are even greater because the targeting is much more accurate and better managed.
- Cdiscount doubles ROI of promotional campaigns
- Allopneus generates 15% more turnover
- Toyota doubles the performance of its lead generation campaign
- The complete guide to personalization and its benefits
- The 4 benefits of retail personalization this peak
- Delivering personalization at scale to unlock $1.7 trillion in new revenue
- With Kameleoon Predict™, your visitors enjoy experiences customized to their purchase intention
- Switching to future mode – why personalization is the answer to successfully managing change
- Accelerating business success post-lockdown with AI-driven personalization
What sectors benefit most from personalization?
Although goals may vary greatly between sectors, personalization benefits all brands.
On an e-commerce website, the goal is to increase conversions and retain customers. Use cases could include adding a personalized upsell block to the product information page, or using targeted promotional actions to convert undecided visitors.
TRAVEL AND TOURISM
On a travel website, conversion takes longer as purchases are significant (holidays, plane tickets). Example use cases might be adding a web-to-call pop-in for an advisor to phone the visitor back, or displaying urgency messages (e.g. this price is only available for three more hours/only four seats left on these dates).
On automotive websites, it is essential to support visitors and help them with this major investment. For example, brands could reorganize the offer carousel on the homepage according to the visitor’s preferred vehicles, or deliver personalized test-drive requests.
BANKING AND INSURANCE
On banking websites, it is important for the customer to receive the same support as if they were in a branch. Use cases could involve personalizing offer pages depending on the customer’s credit score, or reorganizing services based on their needs.
On media websites, visitor engagement is essential to increasing subscriptions and generating return visits. One way of achieving this would be through a personalized pop-in that offers a newsletter subscription, while another might be reorganizing the homepage according to the types of articles the visitor prefers.
- Personalization in E-commerce & Retail: Bring the experience that sales advisers provide in-store to the web
- Personalization in the travel and tourism sector: The travel and leisure experience starts online
- Personalization in the automotive sector: Identify your future buyers online
- Personalized online service and engaged customers
- 4 digital key challenges for financial services - and how experimentation and personalization overcome them
- Personalization in the media sector: relevant content, an engaged audience
- 5 ways AI personalization can increase media revenues
How to put a personalization strategy in place
Kameleoon’s consultants work closely with our clients to help them implement the best personalization strategy to boost their online conversions.
Based on their experience, here are the four steps to follow to implement your strategy:
- Identify your target
- Trigger a personalized action in real time
- Storyboard your personalization experiences
- Analyze your performance
Identify your target
The key to personalizing the digital experience is segmenting and targeting your audience based on your visitor data.
You can use all the website visitor data that you collect as segmentation criteria that you can activate to understand your audience and identify the targets you wish to address.
The data you gather on your visitors can be highly varied. The relevance of your segmentation and targeting therefore depends on the relevance of the criteria you use.
For more precise segmentation, ideally you should be able to cross-reference hot and cold data.
HOT DATA is all the information created from the current visit. We also call this intent data, as it is very revealing of the visitor’s intention when they are on your site. This is the most important data for personalizing the experience in a relevant manner, because it focuses on the actual visitor behavior when they are on your website.
- Behavioral data: pages viewed, browsing journey, history,
- Price and/or product categories viewed, visit frequency, etc.
- Contextual data: day, time, geolocation, weather, etc.
- Technical data: browser, device, source of acquisition, etc.
COLD DATA, on the other hand, is information collected about the visitor in the past. Most of the time, it comes from third-party data solutions, such as your CRM, DMP or offline data and touchpoint. It covers information such as socio-professional categories, gender, age, and purchase history.
- Marketing segmentation: criteria, methods, applications, and examples
- 3 reasons you should geolocate your buyers' journey
- Forget segmentation: modern marketing needs real-time personalization
Trigger a personalized action in real-time
To build relevant browsing journeys for your different segments, you can contextualize all the elements of your website. Here are some tactics for personalized marketing:
Emphasize the most relevant content for your visitor according to their interests or their geolocation. For example, suggest a local restaurant or regional news, etc.
Adapt the visitor’s journey on your website based on their habits, to enhance their experience. For example, highlight the product categories most viewed by the visitor, or re-organize your website’s sections and menus.
Fulfill visitor needs based on the products they’ve viewed, thanks to simple manual rules such as segments. For example, you could create product recommendations based on the visitor’s purchasing power, target certain products only, or use social proof or urgency messages.
Go further than simply addressing consumers by name in your emails with personalized email marketing. Personalize communication based on your visitors’ on-site browsing behavior and interests.
- Big Data: Why the architecture of your tools is essential
- How can an algorithm predict the conversion behavior of new visitors?
- 8 levers to optimize your conversion funnel
- Driving successful ad retargeting with the Kameleoon Conversion Score
Storyboard your personalization experiences
To manage the marketing pressure on your website it is essential to storyboard your personalization experiences. A personalized experience is only relevant if it is seamless for visitors.
Therefore configure multiple elements to be sure that your visit storyboard meets your goals and visitor expectations.
The time ranges during which the personalization is delivered.
Personalizations are a powerful means of communicating a message, but are perceived as intrusive if they appear repeatedly. You have to be able to adjust their frequency of appearance.
Maximum number of appearances (capping): a personalization can be limited by a sales objective - for example, for the first 100 buyers only. You can also cap the number of impressions of a personalization seen by one visitor.
The same visitor can be the target of several of your personalization actions. To avoid any conflict, assign a level of priority to each of them. Generally speaking, if you increase the number of personalizations you have running, you will quickly need to organize them.
You can set a personalization to trigger according to set stages in a specific customer journey.
Analyze your performance
Once you’ve launched your personalization, you then need to monitor and analyze the results based on the goal(s) you’ve allocated.
Before triggering a personalization, it is crucial to correctly set its main goal (primary KPI), but also all the other indicators it potentially impacts (secondary KPIs).
Be clear about your goals:
- Do you want to increase sales?
- Do you want to increase margins by reducing promotional spending?
- Do you want to free up time for your team by eliminating manual tasks?
To gauge performance, compare the primary and secondary KPIs for the population targeted by your personalization with a control group that has not been exposed to it.
Should you A/B test your personalizations?
To measure the success of a personalization, you should retain a control group sample of neutral users who are not exposed to it. This lets you monitor performance of the personalization in real-time, essentially A/B testing it.
In certain specific cases, this approach is not recommended. For example, when you use personalization to target a limited-time sales offer at a given segment, you don’t gain anything from reducing the size of the targeted segment.
- How to manage personalization experience performance in real time
- Taking control of AI personalization with more meaningful and actionable metrics
The essential features of a personalization platform
Choosing your personalization solution is a crucial step in guaranteeing your future success. You must therefore ensure that your solution includes all the essential elements needed to achieve your goals and works within your constraints.
To be able to execute your personalization strategy, ensure that you have a comprehensive solution that enables you to:
- analyze and segment your audience in real time using GDPR-compliant hot and cold data.
- identify visitors thanks to data matching and allows you to create segments according to very specific criteria or by using AI.
- create unique experiences via a graphic editor or a special interface for developers.
- organize your experiences through storyboarding and capping tools.
- analyze your experiences through a dedicated dashboard and collaborative reporting interface.
Above all, it is important to choose a technology provider who can assist you throughout your project.
A comprehensive platform to activate all conversion levers
Consumers are looking for exceptional customer experiences. To help them find these, you must take actions across the entire optimization chain from general usability to specific visitor journeys and needs.
This is why it is crucial to choose a single platform that will allow you to carry out A/B testing experiments and manual and predictive personalization, using one tool, as well as providing the ability to monitor and steer your different KPIs.
It is only by adopting a comprehensive optimization strategy that brands can hope to gain large scale revenue increases.
- The checklist to follow when choosing your personalization platform
- How does a personalization platform should look like
Adopting personalization in your company
Successful personalization requires you to create a culture of experimentation inside your organization. It relies on you analyzing behavior, quantifiable results and experiences, continually questioning and using data to make business decisions.
To achieve this marketing teams are evolving - new roles are being added and organizational structures are being changed to meet new requirements.
The new roles in personalization
The practice of personalization is challenging existing organizational structures, giving rise to new roles within dedicated project management teams. These include:
Optimization specialists and project managers who have solid digital experience. They have watched the sector evolve and understand its challenges. Their role is to structure the strategy and manage the project by coordinating the different roles and resources.
Developers and designers, in charge of the operational aspect of your personalization strategy. The developer handles the integration and the technical aspect of the experiences. As for the designer, they must have a sound UX knowledge in order to create personalized experiences adapted to visitor needs.
What kind of organizational structure should you adopt?
To set up your dedicated personalization team, you can draw inspiration from the three main types of organizations that exist in companies today:
- A centralized structure that drives the personalization strategy for the entire company and prioritizes experiences according to each team’s needs;
- A decentralized structure with personalization experts in each team, enabling brands to run multiple projects simultaneously;
- A hybrid structure with a central personalization unit and local experts in each team.
Read what some major brands have to say about their internal organization structure in our Complete Guide to Personalization Whitepaper.
- CRO: Building the ideal optimization team to improve your digital performance
- How to create an optimization culture
- Building a culture of experimentation to accelerate performance
- Pitching personalization to your boss
- How to launch your first personalization campaign
Move to the next level by taking our personalization training course
With our online training course, you can deepen your knowledge, learn how to deliver personalized experiences to your visitors, successfully segment your targets and build a high-performance personalization strategy.