The Future of Beauty: Makeup Tailored For You.

When it comes to finding the right shade of make up for your skin, I’m confident that I’m not alone in my frustration. Most, if not all, cosmetic brands offer a range of makeup in different shades, catering to both light and dark skins, with different moisture levels. However, this standard range of shades don’t cover all skin tones, which leaves the consumer constantly wondering if they’re leaving home with their face too light or too dark.

Lancôme, one of the pioneers among the cosmetic industry leaders that have shifted their focus to be more customer-centric, are now engaging in personalized products by co-creating with their consumers.

What is it?

The company recently introduced their new product – Le Tient Particulier. It’s a foundation in the shade custom-made for your skin. The consumers face is scanned with a small handheld gadget and the data is then presented to them on a screen with information about their skin. The consumer has the option to select the moisture level of the formula and it’s coverage intensity. The data is sent to a machine that dispenses the formula’s ingredients and blends it together. Within a few minutes, the personalized product is ready. The consumer can test it on their skin and tweak it to their liking. The final product is then saved to the consumers’ custom Complexion ID, which is printed on the bottle with their name, for easy refills. (Lancôme 2017a)

In this experience, the consumers needs are centred and the consumers become part of the value creation process . Lancôme is also able to build stronger relationships with their consumers. Their business model is to offer high quality products that are relevant for all consumers, in addition to meeting their goal of “continually taking science and creativity to new levels” (Lancôme 2017b).

Efficiency Criteria

The joint profitability criteria is met as the product maximizes the joint payoffs for both the consumers and the company. The consumers benefit from obtaining the product designed for their skin, which perfectly meets their needs, thus eliminating the need for them to continue experimenting with different products or settling for one that is less satisfactory. Furthermore, consumers will be able to enjoy the feeling of accomplishment that arise by modifying the product to their preferences (Franke, Schreier & Kaiser 2010).

Even though the company is faced with the investment cost of the machines per store location, the company benefits from the high switching costs consumers experience after purchasing this product. They will likely continue to purchase the product as they’ve found the ideal formula for their skin. In addition, the company further secures consumer loyalty by offering consumers the option to tweak the formula if the consumers become unhappy with it.

The feasibility of the required allocations is also met. While the polity and judiciary dimensions of the institutional environment are not directly related in this case, the social norms dimension is met. Lancôme is a reputable luxury skincare and cosmetics brand, which increases trust in its brand.


Franke, N, Schreier, M & Kaiser, U 2010, ‘The “I Designed It Myself” Effect in Mass Customization’, Management Science, vol. 56, no. 1, pp. 125 – 140, viewed 12 February 2017, <;.

Lancôme 2017a, Le Tient Particulier, Lancôme, viewed 16 February 2017, <;.

Lancôme 2017b, Discover Lancôme, Lancôme, viewed 16 Februrary 2017, <;.



The Impact of Timing and Product Portfolio of Recommendation Systems on Customer Satisfaction

It’s safe to say we have all bought something online. The web has become an important platform over the years for people to obtain information and shop. Why? It’s easy, you can shop whenever, wherever you want and all the information you need is in the product description. Because of rise of e-commerce, personalised recommendation was created to recommend products that meet consumers’ preferences, reduce cognitive efforts, improve user experience, and help purchasing decisions while prompting sales. We’ve all seen companies such as,, and, and even online supermarkets use these recommendation tools. By looking at their consumers browsing history, purchase history and comment history these companies can determine consumer behavioural preferences and recommend consumers the products that they may interest.

The paper written by Yan, Q. et al., looks at the decision-making process of consumers and analyses the mechanisms involved in consumers’ acceptance of these recommendations.  What makes the paper unique is its distinctive assessment of the personalised recommendation system by analysing it from two angles; the recommendation timing and product portfolio. Past papers looked at accuracy and efficiency of recommendation algorithms and their ways to reduce perceived risks, however according to Yan al, a good recommendation system does not only focus on accuracy but also on customer satisfaction which isn’t determined by accuracy. What does determine customer satisfaction is time.

How can time increase customer satisfaction? By recommending products at the right time with the right diversity. According to the preference inconsistency theory, there is a discrepancy of consideration sets in the first and second stage of the decision-making process. In the first stage, when users are browsing, for example for a new pair of jeans, consumers want a lot of choices while in the second stage, before users click submit for purchase, the focus is to minimise the difficulty in decision making and making the right decisions. Too much product choices will cause users cognitive overload, and lower consumer satisfaction. Hence, consumer preferences for recommended products vary in time and the recommended product portfolio and recommendation timing should be consistent with the consumers’ preferences, or it can cause a burden on consumers and decrease consumers’ satisfaction of the system!

What also affects consumer satisfaction is the difference in the type of products recommended in each stage. When consumers browse e-commerce sites, they tend to focus on their own needs and objectives and conduct search on the initial target product and products in the same category. Because consumers tend to focus more on similar products, similar products recommended by the system will be recommended. However, in the second stage, consumers have developed certain awareness and made choices regarding their target product, hence their focus easily moves to products complementary to the target product and consider purchasing other products that are not the target product!

Also, there is a difference in the acceptance of personalised recommendation between practical and hedonic products. Think about the difference of buying dental floss or buying a new television.  The motives for practical products include meeting basic needs and convenience while the motives for hedonic products is based by perceived fun and entertainment. Hence, consumers are likely to have different cognitive and emotional reactions when purchasing these different products. The research shows that consumers who have hedonic products in their consideration set are more susceptible to the systems product recommendations, compared to practical products!

The strength of the research is its further in-depth analysis of the various factors influencing recommendations on consumers. The study takes a different approach compared to past research papers and can be a theoretical basis for e-commerce companies in understanding consumers focus and behaviours at the different stages of the shopping journey. The meticulous understanding can be used to improve customer satisfaction by reducing the cognitive journey and ultimately increase sales! Recommendation systems based on accurate timing and product portfolio are a win-win situation for both the consumer and the retailer!


Tsekouras, D. (2017). Session 2: Personalization & Product Recommendations.

What every marketer should know about hedonic shoppers. (2017). [online] The Rooster Blog. Available at: [Accessed 10 Feb. 2017].

Yan, Q., Zhang, L., Li, Y., Wu, S., Sun, T., Wang, L. and Chen, H. (2016). Effects of product portfolios and recommendation timing in the efficiency of personalized recommendation. Journal of Consumer Behaviour, 15(6), pp.516-526.

UNITED WARDROBE An Infinite Closet in Your Pocket

Imagine you bought a pair of sneakers. After wearing them a few times you realize they don’t fit properly. Even though they are as good as new, you are not able to return them. You could try to resell them online on Facebook or Marktplaats, but you have some uncertainties about safety and security. This is where United Wardrobe comes in: a hip, social and safe fashion platform.

United Wardrobe is an online platform for buying and selling second hand fashion. The key aspects of the platform are safety, sustainability and service. But United Wardrobe is more than just a marketplace platform, it is a community where you can chat with other fashion lovers, follow users and favorite each other’s products. These social functions empower users to become co-creators of value.

How does it work?
A user can create a profile and upload products for sale. The moment a buyer has paid for a product, the seller receives their contact details. As soon as the package has been received, United Wardrobe transfers the money within 14 days to the seller (United Wardrobe, 2017). This relates to what Carson et al. (1999) define as institutional arrangements, the formal and informal rules of exchange created by specific parties to a specific exchange, in this case the exchange of fashion.

The institutional arrangements of United Wardrobe meet three criteria set by Carson et al. (1999). Firstly, they are efficient in a sense that they enable joint profitability and create incentives for users to contribute. Next to this, they are feasible given the characteristics of the exchange of products. Finally, they are achievable in a sense that United Wardrobe has succeeded in growing the platform and community. These institutional arrangements allow United Wardrobe to tackle safety and security issues such as scamming, which no other marketplace platform has succeeded to do.

Users are an important part of United Wardrobe’s business model and enable more creation of value than the company could create on its own. In fact, without its users, the company would not even exist. This is the essence of value co-creation, where new ways are identified to support either the customer’s or the firm’s value-creating process (Saarijärvi et al., 2013). An interesting feature on the website is a page where you can see what the most popular search terms are. This reflects a customer value co-creation mechanism where the firm has refined user data and returned it to the users (Saarijärvi et al., 2013). United Wardrobe has won several prizes with its concept including Dutch Online Retail Experience Award 2015 and the public award of Accenture’s Innovation Awards in 2014.

From my own experience with the platform I can assure you that it is a fun and easy way to sell some clothes. Everyone has clothing at the back of their closet they never wear. A pair of trousers that you might hate another might love, so get up and make that extra money. From an environmental perspective I think this business model is a great step towards a better planet by recycling fashion.

Carson, S. J., Devinney, T. M., Dowling, G. R., & John, G. (1999). Understanding institutional designs within marketing value systems. Journal of Marketing, 115-130.

Saarijärvi, H., Kannan, P. K., & Kuusela, H. (2013). Value co-creation: theoretical approaches and practical implications. European Business Review, 25(1), 6-19.

United Wardrobe (2017) Available at: Accessed on 15/02/2017

Time is money – use Upwork

No matter if you are an entrepreneur, a student, a chef or a CEO – the number of hours you have in a day is fixed. Other than changing the time on your clock, there’s nothing you can do about it. However, you can alter the way in which you spend your time. If time is valuable to you then you are most probably interested in getting the most out of your day. And if that is the case, then you surely must have heard of Upwork. If not, then stop everything you are doing and press play on this video below.

TLDW: Upwork is an online platform for connecting businesses and talent across all industries.

So how does their business model work?

Imagine you have a brilliant idea to start a wristwatch business. You’ve taken care of everything from sales to operations to marketing to HR. However, you still missing a logo. If you have some designer friends, you can ask them to draft something for you. However, you are limited only to those friends and potential referrals & if you ask them to do it for free, you face a risk of their disappointment once they see that your final design isn’t theirs.

Probably not a good idea.

Instead, you sign up on Upwork and create a job post titled “Looking for a graphic designer for my sick new wristwatch project.” After clicking submit, you immediately start receiving requests from multiple freelancers that are fighting for you to accept them. After you choose the one freelancer that fits best in terms of your budget and their expertise, you are ready to proceed with the payment.

Upwork charges both sides of the platform – the client & the freelancer. As a client – you will be charged a 2.75% fee from the total amount that you spend on the freelancer. The freelancer fees start from 5% up to 20% of the total amount.

Not bad, eh? 

Why even use Upwork? 

The system is an efficient and effective tool for both the client and the freelancer. No matter what industry the client or the freelancer come from, they can always utilize this platform effectively due to the powerful matching algorithm.

The clients benefit from:

  • Saving time
  • Saving money
  • Access to a large pool of talent
  • Transparent portfolios of freelancers
  • Efficient tools for tracking progress

while, the freelancers benefit from

  • Additional income
  • Access to a large pool of projects/companies
  • Transparent portfolios of clients

Given that this tool is very user friendly, the investment & costs that both sides of the network incur are quite insignificant:

  • Upwork fees
  • Remote communication (usually the case)
  • Risk of output not meeting standards (usually mitigated)

The relationships created between clients and freelancers work very well, since the platform provides all support for communication (chat & video) for clarification, to hour tracking for monitoring purposes to payment processing for transaction security. By integrating all of these processes, the Institutional Arrangements are managed efficiently.

Since the fundamental idea of this platform is to create new jobs for freelancers, the Institutional Environment is managed depending on the country’s social, political and legal views. A great example of this is the USA, where Donald Trumps policies are aiming to “bring jobs back to America.” This could definitely have an impact on the platform, since the platform connects clients and freelancers globally.

From a societal perspective, there can be similar trouble caused as with the previous example. Although some may perceive outsourcing as the creation of new jobs, others perceive it as having jobs taken away from them as was pointed out in the satirical show South Park.

Politically, Upwork can also cause trouble regarding the discussed issues. Depending on the country & the specific political parties involved, the opinions will vary. If we were to consider an example of republicans versus democrats, we can assume that republicans would be more against the idea of outsourcing jobs abroad while the democrats would support new job creation and potentially even set a minimum wage for freelancers on Upwork within a specific country.



Upwork. (2017, February 16). About us. Retrieved from


“The Filter Bubble: exploring the effects of using recommender systems on content diversity” (Nguyen et al. 2014)

The article chosen addresses the so-called bubble effect identified by Pariser (2012). This bubble effect suggests that by using recommender systems (RS), users are exposed to only a few products that they will like and miss out of many others.  The paper wants to investigate this through understanding the content diversity at an individual level provided by collaborative filtering. It suggests to be the first study observing the effects of this phenomenon on an individual level.

From the study conducted by Lee and Hosanagar (2015) we have understood that there are many opposing views existing in the literature on content diversity on an individual level. Therefore, as the current article claims to be the first one studying this phenomenon on an individual level it is interesting to see how the study has been conducted and what their conclusions are compared to the article of 2015.

The paper addresses very well the debate in regard of the bubble effect: whether recommender systems may be harmful to users. First the behavioral aspects of people who are exposed to similar content and what the effect is on their individual behavior is addressed. As they want to measure the effects on an individual level it is important to recognize what has been found in regard of individual behaviors on content exposure.

For this study, they use the long-term users of MovieLens as they need longitudinal data to draw conclusions on the user’s behavior over time. Two research questions are addressed; 1. Do recommender systems expose users to narrower content over time? 2. How does the experience of users using recommender systems differ of those who do not rely on recommender systems?

The article uses the “tag genome” developed by Vig et al. (2012) to analyze the diversity of the movies that are recommended and consumed (rated). This appears to be a strong measure as it identifies the content of the movie and identifies the similarities content-wise. Multiple articles have used the movie genres (Lee and Hosanagar, 2015) or the ratings given (Adamopoulos and Tuzhilin, 2014) to identify similarities which seems to be less generalizable as the content of the movies still can vary greatly when using these metrics.

The article describes clearly how the findings should be interpreted and addresses multiple questions that have risen based on the findings. This leads to a well-rounded study where the effect of item-item collaborative filtering is exposed. First, the article addresses whether recommender systems expose users to a narrower content over time through comparing the content diversity at the beginning with the content diversity at the end of the user’s rating history. This comparison can show the development of content consumption of the user over time. It has been found that the content diversity of both user groups (using RS or not to rate movies) becomes quite similar over time. Furthermore, it identifies whether using RS reduces the total content-diversity consumed of that user. The conclusion is that users using RS over time consume more diverse content then users ignoring RS. Finally, the experience of the two user groups are evaluated and it is observed that users using RS seem to consume more enjoyable movies based on their ratings given.

As the limitations of the article suggest itself it would be interesting to study the phenomenon in a more experimental setting where the behavior of users can be observed in more detail. This would help in understanding the reasons of the decisions made by the users based on the recommendations. The multiple studies conducted in the field of RS mostly focus on collaborative filtering as this RS is the most commonly used (Lee and Hosanagar, 2015) but research should also focus on other recommender systems to make sure that those used benefit the user the most.



Adamopoulos, P. and Tuzhilin, A., 2014, October. On over-specialization and concentration bias of recommendations: Probabilistic neighborhood selection in collaborative filtering systems. In Proceedings of the 8th ACM Conference on Recommender systems (pp. 153-160). ACM.

Lee, D. and Hosanagar, K., 2015. ‘People Who Liked This Study Also Liked’: An Empirical Investigation of the Impact of Recommender Systems on Sales Diversity.

Nguyen, T.T., Hui, P.M., Harper, F.M., Terveen, L. and Konstan, J.A., 2014, April. Exploring the filter bubble: the effect of using recommender systems on content diversity. In Proceedings of the 23rd international conference on World wide web (pp. 677-686). ACM.

Assessing the impact of recommender agents on on-line consumer unplanned purchase behavior

Recommendation agents (RAs) are often used in modern applications that expose the user to a large array of items (Shani & Gunawardana, 2011). They provide the user with a set of personalized items, tailored to their preferences (Knijnenburg, Willemsen, Gantner, Soncu & Newell, 2012). Hostler, Yoon, Guo, Guimaraes & Forgionne (2012) tested the impact of RAs on online consumer unplanned purchase behavior. Furthermore, customer satisfaction with the website and other variables were included, namely:

  • Product promotion effectiveness: The ability of the RA to recommend products that attract participants’ attention and interest them.
  • Product search effectiveness: The ability of the RA to reduce the extent of product search by providing quick access to relevant information.

Hostler et al. (2012) conducted an experiment in which they created a shopping simulation of ‘online purchase of home movies’. Participants were randomly assigned to a control or treatment group and filled out a pre-test questionnaire in which they rated movies and provided information on their online shopping experience. Second, participants listed the movies they considered purchasing. The control group did this without the RA’s assistance, while the treatment group received personalized recommendations from the RA. Hereafter, subjects filled out a survey on impulse buying, the effectiveness of the product promotion and their satisfaction with the website.


Figure 1. Conceptual model of the results

The results (figure 1) show that the effect of product promotion effectiveness on customer satisfaction is the highest, due to effective product promotion through the use of an RA. Furthermore, customer satisfaction and search effectiveness positively impact unplanned purchases.

In the domain of e-commerce, little research has been conducted on how consumer behavior and website satisfaction are affected by RAs. The findings confirm the importance of using RAs to positively influence online consumers’ purchase behavior and fill a gap in the literature. Furthermore, the findings are also relevant for managers, since they indicate that RAs should be designed in ways to allow greater search effectiveness and satisfaction. Letting users answer more questions about their preferences is one way to increase search effectiveness and might also increase product promotion effectiveness.

A business example that effectively uses RAs is Netflix. Approximately 75% of viewer activity is based on recommendations (Vanderbilt, 2013). Netflix’s recommendations are based on i.e. movies users played/ searched for, including the time, date and device. When it comes to giving recommendations, Netflix aims to do this within 90 seconds, because it knows that after 90 seconds users will abandon the service (O’Reilly, 2016). This shows that search effectiveness is important for Netflix.

To improve this study, marketers should identify other sources for generating customer satisfaction and search effectiveness in the use of RAs. The extent to which RAs and websites are user-friendly might affect these variables. The boundary conditions for the positive effect of search effectiveness on unplanned purchases could also be tested, by examining possible moderators such as product knowledge. Lastly, although this study provides insights in the use of RAs, it is limited to the one product category, namely online purchases of home videos. Since this is a relatively simple product, it would be interesting to see whether the results are generalizable to other product categories, such as complex digital products.


Hostler, R. E., Yoon, V. Y., Guo, Z., Guimaraes, T., & Forgionne, G. (2011). Assessing the impact of recommender agents on on-line consumer unplanned purchase behavior. Information & Management, 48(8), 336-343.

Knijnenburg, B. P., Willemsen, M. C., Gantner, Z., Soncu, H., & Newell, C. (2012). Explaining the user experience of recommender systems. User Modeling and User-Adapted Interaction, 22(4-5), 441-504.

O’Reilly, L. (2016, February 26). Netflix lifted the lid on how the algorithm that recommends you titles to watch actually works. Retrieved February 14, 2017, from

Shani, G., & Gunawardana, A. (2011). Evaluating recommendation systems. In Recommender systems handbook (pp. 257-297). Springer US.

Vanderbilt, T. (2013, July 8). The Science Behind the Netflix Algorithms That Decide What You’ll Watch Next. Retrieved February 14, 2017, from

Managing Consumer Privacy Concerns in Personalization: A Strategic Analysis of Privacy Protection

In the digital age that we are living, one of the major concerns is the protection of our privacy. Research shows that 90% of all online consumers either do not disclose any personal information to companies at all or choose to disclose only to the ones committed to fully protecting their privacy (Taylor, 2003). On the other hand, companies need as much data regarding their customers as possible, in order to be able to provide them with effective personalized product recommendations.

The study of Lee, Ahn et al. delves into this topic by following a very interesting method of research. By implementing game theory, the authors studied the impact of autonomous privacy protection decisions, by firms, on competition, pricing and social welfare. Additionally, this research sheds light on the impact of a regulated environment, regarding privacy protection implementation, on social welfare.

The three main findings:

  1. Asymmetric protection mitigates competition. In simple words, when there are differences between the privacy protection measures that firms implement in any given market, the firm with the strongest privacy protection policy is able to increase its profitability by getting access to a wider pool of consumer data. This simply happens because this firm inspires customers to feel confident to share their personal data with it.
  2. The strategies that firms implement regarding privacy protection should be based on two criteria:
  • Investment cost of protection. This factor introduces the notion that firms in order to implement a privacy protection policy incur some costs such as, infrastructure, personnel and training costs.
  • Size of the personalization scope. This perception regards the pool of the customers for which companies possess personal information and thus are in a position of offering them personalized products or services.
  1. Regulation is socially desirable. According to the research, this holds true since, although the autonomous decisions of firms improve social welfare in general, they redistribute the benefits between firms and customers with firms enjoying the benefits and customers becoming worse-off.

As it can become easily understood, there are many stakeholders when it comes to privacy protection decisions. This research provides a robust foundation regarding the factors that managers should take into account while making decisions concerning their firm’s privacy protection policy. As far as academia is concerned, it connects privacy, in a personalization setting, with equilibria points regarding competition in a market setting. Finally, regulators have one additional source of guarantee that the introduction of privacy protection legislation will be beneficial for society.

In order for all the interested parties to be able to evaluate the findings of this study, it should be underlined, that the authors, in order to calculate the equilibrium in the market, used the notion of firm and customer privacy calculus. This notion advocates that both consumers and firms are perfectly capable of calculating the profits and costs of disclosure of their personal information and the decision of the implementation of a protection strategy respectively. However, this might not always be the case since a lot of biases take place in these processes, as research has already proven.



Sutanto, J, Palme, E, Chuan-Hoo, T, & Chee Wei, P 2013, ‘ADDRESSING THE PERSONALIZATION-PRIVACY PARADOX: AN EMPIRICAL ASSESSMENT FROM A FIELD EXPERIMENT ON SMARTPHONE USERS’, MIS Quarterly, 37, 4, pp. 1141-A5, Business Source Premier, EBSCOhost, viewed 16 February 2017.

Taylor, H. 2003. “Most People Are ‘Privacy Pragmatists’ Who, While Concerned about Privacy, Will Sometimes Trade it Off for Other Benefits,” The Harris Poll #17, Harris Interactive, New York, March 19 (available at http://www.harrisinteractive. com/harris_poll/index.asp? pid=365)

The Visible Hand? Demand Effects of Recommendation Networks in Electronic Markets

When you watch a movie on Netflix, when you listen to music on Spotify, when you watch clips on Youtube, when you search for connections on LinkedIn or when you are shopping online on, it appears often that you come across a sentence like: “you might also like”, “people you may know”, “customers who bought this item, also bought…”. Sometimes this suggestion might interest you and you click on it, but that is not always the case. Why do these recommendations appear and how do companies find the correct ones to recommend? What do these products have in common and how do they influence each other?


This paper of Oestreicher-Singer and Sundararajan (2012) explains these questions by focusing on online product or copurchase networks. In these networks, related products that each have their own network position, are linked to each other. The associations among products, and thus the product’s virtual shelf positions, are visible to the customers through recommendation hyperlinks. The main result states that this visibility of networks can cause a threefold average increase in the influence that complementary products, and thus not only recommended products, have on each other’s demand levels and that it amplifies the shared purchasing of complementary products.

To come to this result, data about the copurchase network of 250,000 books sold on is collected, which is used to test how demand levels are related. The data are tested on three types of possible correlations in demand. Firstly the visible network neighbours with explicit visible hyperlinks. Secondly the complementary products without visible hyperlinks, but with related demand which is controlled for unobserved sources of complementarity that might exist regardless of a visible hyperlink by constructing three alternative sets of complementary products and finally the similar environmental conditions with similar individual or environmental characteristics of the products. Besides these types of correlations, the study found that demand is also affected by the product’s individual characteristics of price, secondary market offers, vintage, in-degree and assortative mixing.

Strength of this study is that a real life setting is used to test the interdependencies, which increases the validity. Downside however is that the study is only about books. For the generalizability, it would be better to also look at other products and services such as movies, cameras or clothing.

The most remarkable outcome, besides the main result, is probably that this visibility has a stronger influence on newer and more popular products, because they ‘use’ the attention of their network position more efficiently. Recently published products are more influenced by neighbouring products, because the effect of observational learning on sales will be smaller when a consumer already has a strong prior idea of a product. Additionally, the conversion rate of recommendations that originate from more popular products is higher and sequentially the same level of total incoming traffic from fewer, more popular sources is associated with higher demand.

This study is important, because as the importance of electronic commerce continues to grow, the ability to control cross-product effects in electronic markets has become a key strategic marketing lever for firms, especially with new and popular products, and that is exactly why you see recommendation sentences when you shop online.



Oestreicher-Singer, G., Sundararajan, A. (2012) The Visible Hand? Demand Effects of Recommendation Networks in Electronic Markets. Management Science 58(11):1963-1981.

Tsekouras, D. (2017) ‘Session 2: Personalization & Product Recommendations, Rotterdam: Rotterdam School of Management (9th of February 2017).

Will the global village fracture into tribes? Recommender systems and their effects on consumer fragmentation

Netflix, Spotify and Amazon and many other companies recommend personalized content to best suit the customers’ preferences and increase consumption or sales (De et al., 2007). Recommender systems are used for different types of media such as movies, books, music, news and television. Moreover, they strongly affect what customers view and buying behavior. Many positive effects of personalization are known such as reduction of information overload, increasing relevance and loyalty (Tsekouras, lecture 2017).

However, personalization may also have drawbacks. As the internet becomes more and more specific to our interests, this “ hyperspecification” may fragment users, reduce shared experience and narrow media consumption. For example, fully personalizing news sites may mean that we no longer see the same news articles. Therefore one could argue that recommender systems will create fragmentation and will cause users to have less in common.
On the other hand, recommenders may have the opposite effect as they share information among customers who otherwise would not have communicated. The paper presents empirical evidence on whether recommender systems fragment or homogenize customers.

In an observational two-group experiment the behavior of the participants before and after recommendations is compared. Recommendations are provided in iTunes and the user experience is personalized through mostly content-based recommendations and a small part collaborative data.

Interestingly, contrary to the author’s expectations, recommendation systems increase purchase similarity (homogeneity among customers) and therefore do not increase fragmentation. Customers tend to purchase more after being exposed to personalized recommendations due to the volume effect, which in turn increases the chance of customers purchasing the same product.
Furthermore, customers buy a more similar mix of products after recommendations due to the product-mix effect. Consequently, customer networks become denser and smaller. However, it is important to note that the recommender system does not recommend the same items to many users but that the diversity of items consumed increases.

Within business, the findings of the paper confirm that recommendation tools can help marketers to increase sales but also alter the product mix customers buy. Off course, these are some of the the reason why numerous retailers (e.g. Netflix, Amazon, Spotify) use these recommendation systems.
However, what I find more interesting about the paper is the application of recommendation systems in the context of society. As we increase to spend time online and recommendation systems become more advanced, I think this is an interesting application. Similarly, I think it would be interesting to research to what extent the proposed homogenization of customers has implications for society and (business) cultures.
Fragmentation may have negative implications for society as when people share little information filter bubbles and echo chambers may exist. However, in business context, I think there are situations in which fragmentation of customers may provide better results than homogeinity . For example in the context of innovation, crowdsourcing and generating new business ideas, a variety of views and opinions may yield to better and more refreshing ideas from different anchors. In this case, it would not be desirable if every customer listens to the same music and reads the same news articles and books, as different point of views are needed. I think for companies who want to generate new ideas from their customers, the homogenization of their customer base might be something to take into account.


De P, Hu YJ, Rahman MS (2010) Technology usage and online sales: An empirical study. Management Sci. 56(11):1930–1945.

Hosanagar, K., Fleder, D., Lee, D., & Buja, A. (2013). Will the global village fracture into tribes? Recommender systems and their effects on consumer fragmentation. Management Science60(4), 805-823.

Tsekouras, D. (10 February 2017), Lecture Customer Centric Digital Commerce, “Personalization & Product Recommendations.

Analysis on an Academic Article: Crowdsourcing Systems on the World-Wide Web

In the article Crowdsourcing Systems
on the World-Wide Web (2011), Doan et al. assess the contemporary business landscape in online crowdsourcing (CS) models. Using a plethora of real-life examples of varying notoriety and a practical, taxonomic structure, the article examines specific types of CS architectures and uncovers the key challenges faced by such systems today. The authors develop an uncharacteristically broad (inclusive) definition for CS which encompasses both direct user contributions like Wikis or idea generators, and passive user value like on social media sites. Compared to other academia on this topic, Doan et al. develop a definition of CS based on the intention of these systems as problem-solving, rather than just the crowd-based method through which they operate. Thus, they define CS systems as ‘[Any system that] enlists a crowd of humans to help solve a problem defined by the system owners.’

While the article contains no scientific testing or experimentation, and offers little in terms of an academic research agenda, it offers some practical value. The authors develop an untypical taxonomy based on nine characteristics of CS systems, and use these to determine system typologies. A distinction is first made between CS systems that use explicit and implicit methods of user collaboration. Traditional platforms like Wikipedia or Mechanical Turk are overt (explicit) about the value users generate, systems like social media sites or ReCaptcha are not. Compared to the other, more academically pervasive dimensions used (e.g. user inputs and system architecture), this initial grouping is unusual as it forces the inclusion of so-called ‘piggybacking’ systems. According to the authors, these are systems in which there are no direct users. Rather, the system uses traces left by external users on other sites (often search engines like Google) to solve their defined problem.

The unconventionally broad definition of CS used by this article lends itself well to a theoretical taxonomy, but it presents problems in the application of such. By including not only implicit systems that lack overt problems or solutions for users to engage in, but ‘piggybacking’ systems that lack users entirely, the tangible concept of CS is made opaque. Even the Internet in its entirety is included in this definition. This makes it hard to categorize real examples concretely, and reduces confidence in any of the actionable recommendations given that pertain to just one category (“does what applies to the internet really apply to my business?”). Additionally, the paper would benefit from modern revision. While the examples are outdated, as is expected from a fast-moving area such as this, they still serve as effective prototypes of their practices today.

In addition to developing a taxonomy on CS system types, the article lastly discusses challenges most often faced by such systems and recommends approaches based on practical examples. While the examples are slightly outdated, as is expected from a fast-moving area such as this, they still serve as effective prototypes of their practices today.

Four distinct challenges are discussed:

  • how to recruit and retain users?
  • what can users do?
  • how to combine their inputs?
  • how to evaluate them?

For each challenge, recommendations are given for assorted CS types using real-life examples – allowing readers an insight into how existing practices have solved or circumvented these challenges in various CS contexts. These challenge recommendations are the crux of this paper’s value as it lacks any direct research or academic agenda of any kind. Regardless, the authors have constructed a useful, albeit not scholastic, lens through which to examine crowdsourcing.

Personalized Online Adverstising Effectiveness: The Interplay of What, When, and Wher

If you go to any website, or online store specifically, your behaviour is tracked. Landing page, time spent, clicks, exit page: you name it, it is tracked. But even when you leave a page, a company does not really leave you: they saw what you clicked on, and based on your browsing behaviour, they retarget you: they show you a (sometimes personalized) advertisement on another channel, hoping you will come back and purchase the product you viewed.

Retargeting can either be done during or after a website visit, and is done based on a customer’s visit. When showing personalized recommendations, for example, it is important to take into account the quality of the recommendation, the level of personalization and the timing. This is what Bleier & Eisenbeiss (2015) looked at: what should they show, when should they show it, and where should they show it.

As with any academic article, past literature is analysed and hypothesis are developed. In order to test the what, when and where of personalized online advertising effectiveness, Bleier & Eisenbeiss conduct two large-scale field experiments and two lab-experiments. The first field experiment looked at the interplay of degree of content personalization(DCP), state, and the time that has passed since the last online store visit, at a large fashion and sports goods retailer, who carries over 30000 products. The second field experiment, conducted at the same retailer, looked at the interplay of placement and personalization. Based on the results from these two field experiments, two lab experiments were designed: one focussing on web browsing in an experiential model, the other focussing on goal-direct web browsing.
Within this paper, thus, many things are studied and confirmed. The papers shows the importance of how to determine the effectiveness of online personalization’s, and which one works best when. When a customer sees a personalized ad right after his/her website visit, the ad becomes more effective. This is mainly because preferences are not constant: they can over time. Thus if you liked a shirt 5 minutes ago, you will mostly still like it now. Thus if a company is able to directly respond to a consumer’s behaviour, the CTR is expected to be higher.

While the effectiveness of recommendations decreases over time, the level of personalization plays a moderating role. This means that high-level personalization in later stages of the decision making process have lower effectiveness, because of changes in customers tastes’ and preferences. The personalized ad is therefore not applicable anymore. Thus, the more personalized an ad, the sooner after a website visit it should be sent. Moderate personalized ads are thus more effective over time, as they take into account these changes in preferences. As visual recommendations are often highly personalized, these type of recommendations are more relevant shortly after a visit. Cross-sell recommendation, which is a more moderate recommendation type, performs better later in time

So what does this all mean? When retargeting customers and showing them personalized ads, it is important to keep in mind how long ago they visited a website. Given that this research was performed at a large fashion/sports retailer, it would be interesting to see whether the same conclusions hold for other settings. What do you think? And when do you consider (personalized) ads target to you most effective?

Bleier, A., & Eisenbeiss, M. (2015). Personalized Online Adverstising Effectiveness: The Interplay of What, When, and Where. Marketing Science, 669-688.

In Managing Teams: GroupMe

Are you spending a spring term abroad in Europe or the States?  Are you just starting your first year of college or graduate school?  For those young people venturing off into a new stage of life, often, studying and meeting new groups of people require being absorbed by a new social platform.  Discussing events or topics that involve everyone’s attention becomes difficult until you introduce GroupMe.  GroupMe is a relatively new chatting application from Microsoft that mixes the best aspects of group chats and social media, with functional tools like calendars, notes and photo galleries.  It makes the entire experience of coordinating nights out, discussing where to go next or discuss possible weekend trips infinitely less complicated.

The app works by first creating profiles of each individual prior to being able to use its functionality.  Once a user profile is created, you can have direct chats or group chats with anyone on the platform, similar to other messenger apps.  Also, like other messenger apps, it stores the names and numbers of your friends that have GroupMe, and creates a pseudo buddy list that you can refer to when creating new chats or groups.

Once you create a chat, the conversation is completely private and can only be shared by someone else inside the group.  As such, everyone in the group chat is an admin and has full rights to everything that is being displayed in the chat – primarily the name of the group chat and the image of the group chat.  Admins do not have control over content that they did not produce, only individual users have content editing privileges over their own content. 

Once everyone is chatting away and making great evening plans or discussing fun events coming up the following weekend, users can get organized by creating calendar events with images, links to facebook, location and directions, as well as, additional notes to inform everyone the price and/ or dress code.  These events then ask everyone in the group if they are going or not, thus creating a guest list for the organizers to utilize.

The fun is enhanced even more with the apps photo sharing feature which in addition to sharing it with the group – the uploader can “Meme” the picture – superimposing text over the picture to add to the laughs!

Monetization of the app comes in two forms.  First, it can use the data from the vast number of subscribers to fine-tune other applications to serve their needs.  Apart from the data, the ability to cross-sell into the platform ancillary services, like mobile skype, photo editors, and organization apps can add significant value to the customers and make substantial profit for the developer.

Overall, the app is a fantastic way of organizing large groups of people aimed at organizing events and sharing good times with each other.  It also benefits from the fact that in order to have a chat with these people, they do not have to be your friend or part of your buddy list, which in other apps can prove to be difficult.  One are of functionality that the app lacks is the ability to share files over the platform.  Only images are allowed to be shared between groups – thus study groups are hurt by this, as presentations and articles have to be shared outside of the app.  As such, this makes the app less than efficient in student groups being able to utilize the app for their basic needs.





Foodfully – No worries about having spoiled food in the back of the fridge ever again

Have you ever encountered a situation when you can’t remember what foods are exactly in your fridge? Do you know people throw away 40% of food they buy at store on average? Foodfully appears as a savior to those who don’t want to throw out food because it becomes spoiled in their fridge and helps them reduce food waste.

What is Foodfully and what can it offer?

Foodfully, developed by a company based in Davis, California, is an app that can help users track their food purchase history and record what’s in their fridges and when the food will expire. The app automatically updates users’ grocery purchases and keeps track of the expiration dates. When the food is likely to go bad, the app will send out notifications to remind their users. And when users are trying to cook with those ingredients, the app will show recipes that prioritize the food that will spoil soonest, which saves the users a lot of time. In addition, Foodfully also partners with several grocery stores like Amazon fresh, Kroger, Safeway to connect their grocery shopping and promotions. By reducing waste and getting more discounts, users of Foodfully get to save more money.


The business model of Foodfully

Foodfully provides food tracking and online recipes for its users, reducing the time and effort they’ll need to record what kinds of food they have in the fridge and when the food is going to spoil. For Foodfully’s business model, there are two types of users, namely free-tier users and paid-tier users. Because the app is completely free for free-tier users, Foodfully makes revenue through sponsored cooking and recipes content. They also sell advertisements that can be seen by all its users to companies that serve food-related products. And for its paid-tier users, the app makes money through linking content like purchases and meals selected to health platforms that require food and nutrition accounting, and integrating with fitness wearables.

Efficiency criteria

By the services it offers, Foodfully is able to create joint profitability with its users. Users can use the app to remind themselves what food is going to waste and reduce the chances of wasting money by throwing away spoiled food. Moreover, by looking at the recipes Foodfully provides, users are also able to save time and effort thinking about what dish they can make with those food. In this case, users gain value by saving time and effort; meanwhile, they give value to the company by providing valuable personal information about what food they buy and what kind of dishes they prefer. On the company’s side, foodfully gives value by providing unique services to those who don’t want to waste food, providing accurate personalized recommendation of recipes and advertisements according to the information users provide on the app. The app also gains value by making money through advertisements and sponsored contents.


Füller, J., Mühlbacher, H., Matzler, K., & Jawecki, G. (2009). Consumer empowerment through internet-based co-creation. Journal of Management Information Systems26(3), 71-102.

McGuire. B. (2016). Automated food tracking – know when your food is going bad and how to cook it. Retrieved from : on 2017/2/15




Recommendation networks and the long tail of e-commerce

Nowadays, we almost can’t imagine online shopping without recommendations systems. Popular electronic commerce websites like Amazon,, and so on all have a section with products they personally recommend to their customers. This is often displayed as: ‘You may also like…’ showing multiple products related to the ones you have recently viewed.

Integrating social networks like Facebook and Instagram into the world of electronic commerce is on the up and can contribute to the personalized recommendation systems of online retailers. In this way, customers get personalized recommendations based on what friends in their networks bought. This makes the less popular products, which customer normally not have looked for, more visible and stimulates consumers to buy products that they normally wouldn’t have found. These products are known as ‘the Long Tail’ products and are often presented as ‘Customers like you also bought…’.

To put it differently, if consumers get e-commerce recommendations based on their networks, the level of awareness for less popular products will increase. This means that the distribution of revenue and demand is influenced and shifts more towards a long tail distribution and away from selling primarily the most popular products. Simply by peer-based recommendations, customers will buy more and different products than they would normally have.


Research done by Oestreicher-Singer & Sundararjan (2012), investigates the impact of peer-based recommendations on the demand and revenue distribution. They research the influence of network-based recommendations on the online sales of 250.000 books from online retailer The research shows that by recommending books based on what friends in customers’ networks bought, the distribution of demand and revenue is highly influenced. The researchers focused on the top 20% most popular and top 20% most unpopular products.

Categories of unpopular books that were displayed based on peer-recommendations experienced a 50% increase in revenue whilst the commonly unpopular books experienced a 15% decrease in revenue. This meant that the unpopular books suddenly became more visible to customers which led to an 50% increase in sales.

That all sounds quite impressive, but one could not say that this 50% increase was only caused by the visibility of products through recommendations. Different other contemporary factors also contribute to the redistribution of demand and revenue of consumers. Lower search costs and higher product variety for instance, have a great influence on the long tail of e-commerce.


All things considered, e-commerce is highly influenced by the power of social networks. The influence of recommendation networks positively affects to phenomenon of the long tail demand. Selling less of more rather than more of less is going to characterize the e-commerce demand curve in the future. The implementation of ‘what other customers like you bought’ will continue to impact our online shopping behavior. If companies implement the right recommendation systems to influence consumer demand, the opportunities are endless!



Anderson, C. 2006. The Long Tail: Why the future of Business Is Selling Less of More. Hyperion Press.


Brynjolfsson, E., Hu, Y., and Smith, M.D. 2006. From Nichees to Riches: Anotomy of the Long Tail. Management Review 47(4) 67-71.


Oestreicher-Singer, G., & Sundararajan, A. (2012). Recommendation networks and the long tail of electronic commerce. MIS Quarterly

“Buy a present for my wife” said Jan to the phone

This year St. Valentine’s Day caught millions of men by surprise, again, leaving them wondering what present to buy for their partners. What if somebody or something could perform this burdensome task in a timely manner? There might be a solution…


Viv is an intelligent personal assistant introduced to the market on May 9, 216 and acquired by Samsung in October 2016. Similar products such as Siri, Google Now, Microsoft’s Cortana and Amazon’s Alexa can perform some basic tasks but nothing beyond the tasks they’ve been programmed to do. Due to artificial intelligence, Viv can generate code by itself and learn about the world as it gets exposed to more user requests and new information.

This makes it by no means a universal product. Viv is expected to learn and store information about every user, and learn with time how to serve him or her personally. For example, if the owner asks: “I need to buy a present for my life for St. Valentine’s Day”, Viv should be able to predict what a suitable present would be or perhaps book a table for two at a fancy restaurant downtown.

Continue reading “Buy a present for my wife” said Jan to the phone

Amazon Go: Invisible Interaction Yet Visible Personalization


The IT giant Amazon has been extensively testing its first smart offline store “Amazon Go” in its hometown Seattle, Washington since its exposure at the end of last year (González, 2016). This concept store has gained significant popularity as it claims to have no cashiers or checkout lines thanks to computer vision, deep learning algorithms and sensor fusion.

Continue reading Amazon Go: Invisible Interaction Yet Visible Personalization