Paris Innovation Review – In the early 2000s, before the birth of web 2.0 that was going to transform the user into a full actor of the online experience, we saw the emergence of the “personalization” trend. What was it all about, back then?
Ahlem Abidi-Barthe – When I first conducted my investigation in 2004, brands offering customizable products were spreading all over the web. Mass customization was developed in the early 1990s and appeared to have a promising future with the rise of online business. We imagined consumers with the best of both worlds: the effectiveness and prices of industrial production and mass consumption, but also the freedom to choose, the opportunity to escape from standards, to customize everyday objects.
The pioneer Land’s End offered customers the possibility to order customizable pants and shirts online: clothes were made to order based on a large number of options and varieties of design. In another field, computer geeks also started selecting their ideal computer, at a lower cost, by buying and assembling separate parts themselves. Geeks believed, and made us believe, that we were on the verge of a mass movement that would revolutionize the industry...
What happened ?
Well, that movement did occur, but it took a very different direction. In the field of consumer goods, mass customization has not disappeared, but it remains relatively marginal. Either the right business models haven’t been found, or brands grew tired, or consumers simply weren’t interested. Most brands continue to offer standard products. Apart from T-shirts or shirts, mass customization concerns mainly gadgets such as mobile phones bodies – nothing really significant, cost-wise. How can a concept as promising and brilliant as MyEvian become profitable when water bottles cost up to €5 per bottle? However, we witnessed the rise of a number of niches such as shirts, perfumes or bags and accessories; the appearance of actors specialized in tailor-made, with a high added value: in this regard, the luxury sector could experience significant disruptions. Furthermore, suppliers or distributors of blinds, melamine panels, curtains, offer today, almost systematically, the possibility of ordering custom-made parts. Finally, 3D printing paves the way to extreme personalization, with the creation and production of absolutely unique items. All these developments offer interesting ideas but these do not match with our view on mass customization in the early 2000s.
However, the personalization of the relationship with Internet users has, in turn, evolved considerably in the direction of a greater personalization of the relationship. But one should not confuse this personalized relationship with the customization of products.
Can you clarify this difference?
Let us start with what brings these concepts together: they provide unique experiences to customers of a commercial site or more broadly, to users of an online service. For example, when you set up Google News (by selecting or discarding a theme, focusing on another), you are “customizing.” When Amazon or Voyages-SNCF track your preferences and offer choices that fit with your profile, it’s called personalization. When you set your app features yourself, it’s customization.
The difference between personalization and customization is the level of participation of the user. While the customization of the content of the website, for example, gives the user the ability to actively set the configuration of information on the site, customization gives the user a more passive role: the site assumes the filtering function. Customization is more about the products (bags, cars, clothing, etc.), whereas personalization is about the relationship with the customer.
No need to beat about the bush: customization of products/services by users is not a dominant trend. It continues for free services provided by large platforms (for which, as the saying goes, when it’s free, you are the product), with different philosophies from one platform to another.
Personalization made tremendous progress over the last fifteen years, mainly because of the time we spend online. In a context of social networks and Big Data, we inevitably leave more marks. In other words, we produce more encrypted data but also text, data, photos, videos, etc. More recently, with geolocation, the personalization of relationships and services is becoming even more accurate. Take one example: the weather option in the latest app of La Redoute, a French mail-order retailer. If I am geolocated in a beach town like Biarritz, in August, a push notification instantly offers me swimsuits.
Fifteen years ago, we saw in personalization a crucial tool in building customer loyalty. Is it still the case?
Yes and no. Yes, in the sense that it fits into the broader concept of user experience, which is crucial and proves highly discriminatory. We’ll come back to this later. No, because loyalty is based on other criteria, such as trust, site reputation, quality of the after-sales service, depth of the offer, etc. No, ultimately, because it’s a double-edged tool, and that’s what my study over this period revealed.
What were you conclusions?
I spotted a gap between the discourse of theoreticians and the perception of consumers. Following the seminal 1997 book by Peppers and Rogers (1997), Entreprise One-To-One: Tools For Competing in The Interactive Age, several theorists presented personalization and customization as important precursors of loyalty to merchants websites. The reasoning was the following: loyalty requires a learning relationship between the company and its customer, a relationship that can satisfy more and more precisely the customer’s individual needs. When faced with a variety of clients, a retention strategy needs to personalize its actions and individualize its offers, based on the individual and different needs of everyone.
On paper, it made sense, and professionals showed a genuine enthusiasm by investing massively in CRM (Customer Relationship Management) tools, at the heart of which stands the personalization of customer relations. What do consumers think about it? Weren’t theorists and practitioners trying to answer them when they advocated the need for tailor-made and personalized solutions in order to improve customer retention on merchant websites? That’s precisely what I wanted to check.
First of all, it should be noted that my sample is only based on regular Internet buyers, because I was interested in loyalty to websites and long-term relations (and not in purchase intention or satisfaction). In 2004, these regular buyers were far from forming a representative sample of the population of Internet users.
The theory opened on two possibilities: an explicit personalization (or customization), performed by the client and an implicit personalization, managed by the merchant. We saw what explicit personalization was about. Today, it is confined to a few, very specific areas. Back then, one could already see the low interest of users for these parametric tools, which were perceived as useful... but that nobody used!
The reasons that pushed to buy on the Internet were the practical side of temporal and spatial accessibility (“I sign and I buy when I want”), as well as avoidance of the crowd; the ability to compare prices, of looking at a product with absolute peace of mind, without suffering the advice and opinions of professional salesmen; the effectiveness of searching by keywords. No trace of customization, here.
Why were they sticking to a site? The most important factors were, in order of importance: the after-sales service, prices, scope of offer, the ergonomics and ease of navigation and last (!), personalization (help to choose, advices, etc.).
Customers of merchant websites spontaneously spoke of personalization as a crucial factor of loyalty, but they placed it last. Obviously, we were in 2004 and CRM was still embryonic. But distrust already appeared under certain forms: for consumers in 2004, personalization was mostly synonymous of spamming, junk mail. Not only because they overflowed mailboxes, but because, as one interviewee said: “I don’t like being a target and I do not like being put in a category...”
The negative attitude towards personalization was also linked to negative consumer experiences with companies whose practices and algorithms were not fully developed: “In general, it has nothing to do... they say, a customer who bought this also bought that... but what they offer doesn’t interest me at all.”
Apart from spam and negative experiences, some forms of personalization pleased consumers: being recognized by the site, at the time, was seen as something nice. Similarly, personalization in the form of advice, based on questions directing consumers towards a particular product, was much appreciated.
But on the whole, the perceived benefits were quite low. I could summarize the spirit of these answers as follows: people were quite reluctant to the idea of being “personalized” automatically by a machine. At the time, I was drawing the conclusion that consumers wanted to have control.
For consumers, perceived control was undoubtedly the most important variable regarding customization. But I also identified two other variables that conditioned the consumers’ acceptance of personalization and customization: the evolving nature of personalization and moderate intensity of customization.
The importance of perceived control reflected the fact that on the Internet, consumers don’t like losing control or feeling manipulated. They want to have the choice: the form of personalization, its intensity, searching on the Internet when they want, being a target when they wish to be one and only for that time.
The qualitative study also revealed another result: the use of any form of personalization and customization depends on the consumer’s expertise and involvement in the product category. Interviewees seemed to support customization on product categories for which they had little or no expertise and no strong involvement.
Finally, in 2004, the issue of control over private data kicked in: answering to unrelated questions or without direct utility was already taken badly.
You also were talking about the evolving nature of personalization.
Yes. Among the negative experiences of consumers, there was the non evolutionary character of personalization practiced by some merchant sites: after acquiring a gift, we found ourselves assailed with proposals that correspond to this gift, but not to own’s own tastes. Collaborative filtering solutions developed at the time assumed that consumer needs are static and can be targeted based on the static needs of consumers with the same profile. Reality is more complex and consumers have different needs according to different situations: already at the time, the need for a situational and scalable personalization was already perceived.
Along the same idea, the qualitative study also revealed another very important variable for the consumer, personalization must have a moderate intensity: “More is not better”! Consumers evoked the need for intelligent personalization: little, but well.
This call for moderation, especially against the first attempts of personalized pricing, is well illustrated by the misadventure of an Internet user who caught Amazon.com in an act of dynamic pricing. After wiping all personal data from his computer, data that defined him as a loyal customer of the site, he noticed that the price of a DVD, previously offered at 24.49USD, dropped to 22.74USD!
Geolocation via IP address also provoked unease among Internet users, who deplored an intrusion into their privacy, a Big Brother aspect. But they appreciated being offered a hotel in the destination city of train ticket they were ordering.
A qualitative study of Zeithaml, Parasuraman and Malhotra published in 2000 went along the same lines. Indeed, in standard services (person to person), the perceived service quality is based on “more is better”: the relation between key dimensions of service quality and its comprehensive evaluation by the consumer is more or less linear (the more empathy, reliability, etc. the more consumers assess positively the quality of service). However, on the Internet, this relationship seems to be rather curvilinear: “more is not better.” For many attributes of the quality of service on the Internet, including personalization, consumers seem to have reference points above and below which the perceived quality of service decreases.
All of this is further complicated insofar as the ideal point of personalization varies according to the consumers: while some may appreciate a company knowing them better once they have provided detailed information about their needs and habits, others will find it intrusive.
Many of the points you raised remain relevant in 2016. According to you, what would be the main developments since then, and which of them were considered as surprises?
Let’s start with the surprises. For me, it’s the fact that there are still many hitches in the promises of personalized offers and experiences made by businesses to their customers, despite technological advances, especially in CRM tools and Big Data. These offers continue to annoy consumers and discredit brands, especially when they “brag” and tend to rely on the knowledge of individual offers, just for you. I will always remember this mail I received from Viadeo Career, entitled: “Ahlem, here are your personalized career opportunities.” This so-called personalization was based on a single key word in my profile: “manager.” As consequence, I was offered a tailor-made job… as “manager of butcher’s shop in Courbevoie"!
In reality, human intelligence needs to play a full-part role, in addition to artificial intelligence.
According to me, these changes come from the combination of both practices: the personalization of relationships on one side and the customization of products on the other, in order to build a global experience that reenchants the customer’s act of buying. My doctoral thesis focuses on this issue: the customization of the offer and personalization of relationships as two complementary strategies in building customer loyalty on a merchant website.