Read mode
When Krzysztof Franaszek wanted to better understand the quality of personalized advertising in 2021, he conducted an interesting study using his tool Adalytics (a Chrome extension).
He sought to answer how effectively targeted advertising works.
In some previous posts, we wrote that 83% of users are willing to share their data in exchange for more relevant content (ads).
But do we actually get relevant content in return?
Looking through the lens of our own created data, the responsibility for appropriately tailored content lies entirely with the owner of that data. What about the situation when we trust the parameters and tags offered by third-party data providers or ad networks?
To better understand the answer to this question (written a few lines below), let’s first look at what the “rental” of third-party data, offered by various providers worldwide, looks like.
Lou Montulli, while working at Netscape in 1994, created web cookies to enhance the user experience and, of course, to monetize websites. He probably didn't think he was opening Pandora's box then, but he admits today that some things could have been thought through more thoroughly.
An example of accepting cookies on an EU website about cookies?
Cookies enable websites to remember a user's (your) preferences, which on the next visit, makes the web experience easier and better (e.g., showing relevant content, ads, offer).
Years of using (abusing) cookies have led to the regulation of their use (GDPR legislation). Still, more and more users acknowledge that some content pushed by ad networks is odd, unusual, and sometimes even a bit creepy. Why?
So, how well does targeted advertising actually work?
Krzysztof’s analysis describes in detail the results of a pilot project. For audience research, he used a convenience sample of 25 volunteers.
The small sample size of the study is not statistically representative but is intended as proof of concept for the types of insights that could be obtained with more extensive statistical processing if there were more participants.
For the study, Franaszek tracked the browsing behaviour of 25 volunteers from around the world (OPEC countries) for two weeks. At the end of the period, some received surveys to assess whether the ads shown to them were relevant.
Insights were obtained from the clickthrough URL after clicking on an ad, often revealing details about why someone was shown that particular ad. For example, with interest tags like “traveller”, “gamer”, or other demographic parameters like “women”, “men”…
Although the sample size was not large enough to make definitive statements, the results illustrate the pitfalls of targeted marketing - and the potentially dubious data that advertisers rely on.
90.5% of ads for the Merino shoe brand were targeted at the wrong gender. This means men regularly saw ads for women's shoes and vice versa.
The National Rifle Association (NRA) “repeatedly” served ads to two people it thought were “luxury vehicle enthusiasts”, even though neither person had ever indicated an interest in luxury cars (or firearms).
One user was shown an ad for the Saatva mattress brand, labelling the user as being “deeper in the sales funnel” because they searched for the brand. However, the user only searched for the brand because they received the mattress as a gift and just wanted to check its dimensions.
There’s no simple answer to the posed question, but it’s clear that caution and careful checking of ad networks (and their promises) are appropriate, if not mandatory.
So, how do we equip ourselves with knowledge and appropriate data to minimize the dilemmas described above, or to ensure that power and knowledge are predominantly in our domain, and not just a result of 'trust' (and the sales skills of ad networks)?
Best to start with understanding the types of data.
First-party data are data collected by the specific website that the user visits. It can track behavior as long as the user is on their website and includes clicks made, behavior such as scrolling, hovering the mouse over a part of the site, and time spent on different parts of the website. As soon as the user leaves the website, the first-party cookie no longer follows their behavior.
These data can be used by the company to adjust the user experience, communication, website appearance... on their website. The key advantage of first-party data is its uniqueness to your brand and includes user interactions with your content. Another major advantage of first-party data is that they are free and yours.
The downside of first-party data is that they contain very little general data about user behavior, making it difficult to get a comprehensive view of user interests and engagement.
As a result, it's harder to adapt (personalize) the website to different users based on their behavior.
First-party data are data collected by the specific website that the user visits. It can track behavior as long as the user is on their website and includes clicks made, behavior such as scrolling, hovering the mouse over a part of the site, and time spent on different parts of the website.
Second-party data, or purchased data, can be best described as the first-party data of another brand (company). They are usually bought from third parties and include everything from demographic data to email address lists. The “exchange” often happens in joint co-branding campaigns.
The advantage of these data is that they can enrich your own data base and give you a more complete profile of the user.
The biggest drawback of second-party data is that they are usually quite expensive.
Second-party data, or purchased data, can be best described as the first-party data of another brand (company).
With the expansion of online sales and consequently advertising, the use of other data, usually provided by ad networks and social networks, has logically also expanded. These data are collected, organized, analyzed, contextualized, and sold.
Today, third-party data - primarily due to legislation - are mostly anonymized.
Third-party data are user data collected across different web domains. It is the most comprehensive set of user data, accumulated from various web interactions and domains, with the aim of profiling the user as accurately as possible.
As consumer awareness of online data privacy increases, this form of data collection becomes increasingly problematic.
The main advantage of third-party data is their rich array of parameters and broad historical insight into individual user behavior, browsing, and activity.
One of the main reservations about using third-party data has already been mentioned in the first paragraph, but another important characteristic should not be overlooked. These are not your data. You only rent/buy these data from third-party providers (Facebook, Google, LinkedIn, ad networks…) and thus agree to their rules of the game (and prices).
Third-party data are user data collected across different web domains. It is the most comprehensive set of user data, accumulated from various web interactions and domains, to profile the user as accurately as possible.
It’s all fun & games until the providers of large user systems decide they will no longer allow data recording by third-party cookies (Apple).
What then?
Part of the answer lies in own acquired, or zero-party data. Yes, zero-party or self-acquired data, is an upgrade of own first-party data.
We obtain them simply by asking customers about their desires and needs, instead of just (only) tracking their online behavior.
The simplest way to understand the acquisition of zero-party data is to think about what questions a salesperson in a physical store would ask a customer.
For example, what brand of shoes are you looking for, what style do you like, what colors did you have in mind, etc.
This method enables an honest, clear, and two-way relationship with users. You openly ask them which data they want to share with you and present them with the benefits of sharing such information.
To provide a customized experience for your customers, there's no need for more crawling or spying.
Part of the answer lies in own acquired, or zero-party data. Yes, zero-party or self-acquired data, is an upgrade of own first-party data.
Zero-party data are an upgrade of first-party data. We obtain them simply by asking customers about their desires and needs, instead of just (only) tracking their online behavior.
Comparison of some key factors between Zero and First party data
Zero-party data creates User's personal choices (preferences), while First-party data are Data about behaviour or purchase.
Zero-party data is acquired consciously and with demonstrated interest, while First-party data is acquired unnoticeable by the company
Zero-party data are Predictive – with the desire to improve future user/purchase experience. First-party data are historical, and tied to past user behaviour.
With Zero-party data we respond to the desires and explicitly demonstrated preferences of users, while First-party only allows us to assume further steps based on past data.
The user expects a better user and purchase experience in return when providing Zero-party data, while at first-party data, the user is unaware of the collection and may receive inappropriate (creepy) follow-up content.
The sexiest thing with zero-party data is, that users update them themselves in the desire for a better experience. While users have little to no influence on first-party data updates.
To acquire zero-party data, several types of tools and solutions allow for quick, easy, and content-rich data collection that will help better understand your users.
Tools with which you can successfully and in large quantities collect zero-party data are:
We have created a lot of posts and excellent practical examples about recommendation quizzes on our blog. Both domestic and foreign.
Quizzes with product recommendations sell for you. They act as a personal salesperson in an online store, guiding their customers from start to checkout, and helping them find products that best meet their needs. They do this in a fun way, allowing the customer to interact with your company and encouraging them to complete the quiz and make a purchase.
Once you collect enough quality data, you can start personalizing in the online store. This has been proven to increase conversion rates, improve the user experience, and make the brand more personal. How to approach it correctly and which tools to use for the best results?
The purpose of interactive quizzes (segmentation, personality, knowledge quiz) is not to show the ideal product, but to get to know the user better. What are their preferences, what they don't like, how they think, and how they respond to fun situations.
Marketing quizzes are one of the most popular interactive online contents, which you have surely tried yourself. They are fun, simple, and just challenging enough that we want to find out how well we do.
An example of an engaging quiz, where the user is greeted at the end with a thank you for participating with a gift - a unique discount code, just for quiz players.
Or create a fun segmentation quiz that can improve your advertising communication, sign up more users for your newsletters, and supplement their user profile with new data.
The survey probably doesn’t need a deep explanation. Its primary purpose is to obtain information from the user. As a good example, I can highlight the gardening quiz we created in collaboration with the Skaza brand.
In the fun gardening quiz, where we challenge users with knowledge about gardening, crops, composting, and tools, we asked users an additional simple and straightforward survey question after the fourth and eighth question (the entire quiz has 11 questions).
We thus enriched the thousands of users who played the quiz with the additional information of whether they live in a house or an apartment. Depending on the answer to the first survey question, we adjusted the following questions accordingly. If the user answered the survey question with answer A, they received the next question A1 (Do you have a garden at home?) or, if they answered with B, question B1 (Do you separate organic waste?).
In this way, the client can adjust the feedback, meaningfully segment users in their email client, and create customized content for each segment of the sales funnel.
If you are interested in how many responses we collected in the survey and what the average user drop-off is when we offer them a survey in the quiz, take a look at the case study here.
Interactive forms can also be an excellent space for pleasant and creative communication, enriching data from users who show interest in contacting your store.
For example, we designed blog feedback on our new website in this way. We directly ask the user whether they liked the post. Depending on the first answer, the entire subsequent communication is also adapted. In the case of a positive response, there are more questions and we also ask the user if they want to subscribe to the newsletter, what their name is, this attribute [name], and in the last step, we use it for a personalized address and a question about the company they come from.
Example of an interactive form with which we get user feedback on blog post content.
We have user data, what now? There are a lot of information, tactics, and approaches in the use of Zero-party data. So much so that this section will soon get its own post, where we will look in depth at what we can do with such data.
The most basic, popular, and widespread ways to use your own acquired data are certainly:
Segmenting users into interest groups and tailored communication through various channels (email, Facebook, Instagram, ...)
Customized content on the website (banners, pop-up notifications, promotions, discount codes, ...)
Courses and education related to a specific interest group
Other sales benefits and activities (free shipping for a specific segment, weekend promotion, referral mechanism ...)
We wrote more in-depth about customizing content based on zero-party data in a blog post about personalizing content in an online store.
I have probably listed more than enough reasons for serious consideration of upgrading your own data enrichment strategy. Now it's up to you to translate the information you found in this post into your system. To your brand. To your users.
If you need help creating any of the above-listed solutions, with which you will get to know your users better and honestly ask them about their opinions, interests, preferences, and wishes, we are happy to help.