The perils of personalization – and how to avoid them
Increasingly personalization is essential for all companies aiming to improve the visitor experience and increase conversion rates.
According to Gartner’s Making Personalization Pay study, 67% of marketers have launched personalization efforts, allocating an average of 14% of their budgets to the capability.
However, not all companies have the same level of knowledge and maturity, and it is easy to make mistakes in terms of technology, approach, culture or data use.
To help, we talked to our consultants who work closely with our clients to develop and roll out their personalization strategies. Together, we have identified six pitfalls to avoid when starting out in personalization.
1 Pitfall 1: personalizing without a strategy
A personalization strategy should not be driven by a one-off need (“I want to personalize the slider on my homepage”) but by a carefully considered strategy with specific goals (“I want to increase visitor conversions/loyalty across all web journeys”).
This means it’s not enough to just personalize your emails or your homepage. You need to have an overall objective for your experience that personalization will help you deliver.
For French ecommerce leader Cdiscount this means using personalization as a strategic lever that can cover the entire conversion journey:
The solution: think strategic
Start by modeling existing visitor journeys to identify friction points and opportunities for improvement.
This will help you decide which personalization actions will deliver high added value on your website, improving the overall visitor experience and providing real results in terms of ROI and margins.
2 Pitfall 2: personalizing without a roadmap
When you get good results, it can be tempting to set up multiple personalizations on your website at once without creating an overall roadmap.
This can undermine the whole point of personalization, which is to provide visitors with quality experiences. If you offer the same visitor too many personalized experiences, this may have the opposite effect and spoil the overall experience.
Additionally, if you launch too many personalized actions at the same time, this can also compromise results. Two personalized actions launched simultaneously without prioritizing or storyboarding could turn out to be inefficient if they conflict, or may create too much marketing pressure on the targeted visitors.
The solution: draw up a roadmap and follow it
Before launching your first experiment, you should establish a roadmap to prioritize your experiments and manage marketing activities on your website.
This is both vital and often overlooked – Gartner found that just 29% of marketing leaders had a dedicated personalization roadmap to help them spot opportunities and improvements.
Using your roadmap, you can identify and differentiate between actions that are absolutely necessary for optimizing performance, sporadic actions (promotions) or regular, calendar-based actions (such as around Christmas or school holidays).
This ensures that each experiment is launched in the right conditions and is part of a general strategy aimed at achieving goals set in advance.
3 Pitfall 3: personalizing with the wrong tools
In a recent report, Gartner identified three main difficulties marketers face when choosing a technology partner that will help them develop their personalization program:
Lack of strategy
This is the first pitfall we discussed. Before investing in a personalization solution, you must decide on a strategy. What personalization use cases do you want to implement? Does your industry have any specific requirements?
Difficulty understanding the technical details of technology partners
It is essential to really understand the technology and applications of your personalization platform if you want to get optimal use from it.
Lack of the right in-house structure
If you want to use your personalization solution properly, you need to recruit or train the right people in your company.
These three factors can lead marketers to pick a solution that doesn’t match their needs, negatively impacting their entire optimization strategy.
The solution: choose the right technology partner
Start by identifying your needs in order to understand which functionalities are essential (such as server-side vs client-side, advanced targeting, segment configuration or integrations.)
To learn more read our blog “How to choose your personalization solution.” Based on Gartner’s Magic Quadrant, this lists the essential criteria you should take into account when making your buying decision.
Additionally, choose your solution with a view to the long-term, in order to anticipate future needs. Optimization strategies often start with A/B tests to improve the user-friendliness of a website; personalization is then added to enhance the experience at the segment level while AI-driven personalization is then deployed to deliver a tailored, individual experience.
Taking your future needs into account right from the start makes sense, meaning you can choose an all-in-one tool for your entire optimization strategy.
4 Pitfall 4: personalizing solely on a single channel
Today, 73% of your customers use multiple channels before converting, including apps, the web and social media.
Given that the aim of personalization is to offer a unique experience adapted to visitors’ needs, it is therefore counterproductive to just personalize a single channel as it will not cover their entire purchasing journey.
For example, a visitor who has already run searches on your website should be able to easily continue with these when pursuing their journey at a later stage, even if on a different channel or device.
The solution: develop an omnichannel personalization strategy
To guarantee a quality experience throughout the customer journey, your personalization strategy must integrate all available channels into your personalization strategy.
This requires two things:
- Successfully streamlining visitor journeys to offer a similar experience on all channels.
- Successfully triggering synchronized personalization actions on all channels.
Omnichannel must therefore be a part of your strategy from the outset and an essential criteria when choosing your personalization solution.
French retailer Auchan has rolled out personalization campaigns across all channels with Kameleoon.
By offering personalized promotions on every channel, Auchan has seen an 18% increase in sales amongst targeted segments.
Find out more in this success story.
5 Pitfall 5: not taking advantage of real-time
Personalization aims to offer an experience tailored to visitor needs – and data is essential to understand what they are looking for.
Collecting and analyzing visitor data is therefore the cornerstone of an effective personalization strategy.
That means that when personalizing, it is a mistake to only use historical, cold data as that means ignoring the real-time needs of visitors.
While you can personalize using cold criteria such as most recent orders, it doesn’t take into account the current needs of the visitor. Additionally if you rely on this cold data it is impossible to personalize the experience for visitors who haven’t bought from you online before.
Let’s take an example: a customer visits a website selling electrical goods where they bought a washing machine two weeks earlier. This time, they are looking for a present for a friend whose birthday is coming up; they’d like to buy them juicer.
If you used a criteria such as previous purchases as a base, you would logically suggest a washing machine to this customer. However, this isn’t what they want this time.
The solution: rely on hot data and on a technology partner whose architecture is built for real-time
If you decide to take the real-time personalization route, it is essential that you choose a solution that can handle those open-source technologies (Hadoop, Kafka, Spark, ElasticSearch, etc.) that are able to provide a stable Big Data architecture that can feed data into your back office in real-time.
Rather than cold data, you should use hot data relating to the behavior of visitors during their current visit (such as pages visited, time spent, items added to cart /favorites, etc.). This is the most representative of visitors’ actual purchase intentions at a given moment.
You can take this approach further by using AI-driven personalization, which, with the help of predictive algorithms, enables you to determine the conversion probability of a visitor in real-time and to automatically trigger actions according to this probability.
6 Pitfall 6: personalizing without testing
Optimization is based on learning, and the same applies to personalization. Just because an experiment works for a certain target, or for a defined period of time, doesn’t mean it will always work for everyone.
While you need to create an initial hypothesis based on data analysis, this is not enough. You need to be able to analyze and measure the relevance of the personalization and improve your strategy based on the results.
Moreover, personalization relies on the behavior of visitors and their needs, which evolve. So your personalizations need to be continuously tested to ensure that they are still relevant and adapted to customer’s behavior.
The solution: test and learn with A/B testing
When launching a personalization experiment, you must constantly monitor the results and analyze them in-depth using advanced reporting tools.
It is vital to understand why a personalization is working well in order to learn from it and adapt your strategy and future personalizations.
It is also good practice to A/B test your personalizations. Compare a sample of your visitors exposed to the personalization with a control group that does not experience it. To enable this work with a technology partner that includes A/B testing within its platform.