Do you think using data mining as a black box tool is justifiable?

These users [marketers] either know nothing about the techniques of data mining or do not need to know anything about data mining to reap its benefits“. I remember following a doctoral course on data mining where the teacher, a professor in machine learning, was claiming that people should not use data mining as a black box tool. If this is the case, the so called garbage in, garbage out situation is likely to happen. Do you think using data mining as a black box tool is justifiable? What is your mind about that?

Comment of Erik:

Being one of the authors of the mentionned book, I thought I would try to expose the other side of the coin… it is no fun when everybody agrees on a blog ;-)
At the last KDD conference, there was a panel organized trying to understand why data mining is not a multi-billion dollars business as, for example, it is the case with Business Intelligence (dealing with reports and OLAP: what I call the ‘low end’ of analytics, or human powered analytics by opposition to the maths powered analytics).
It was striking to see that almost all data mining experts that were present (and there is a lot in KDD) were claiming that you do need statisticians or data miners to do data mining, and KXEN representative (Rob Cooley) was almost the ony one claiming that automated data mining is possible. So, the point of view of the people participating to this blog is ‘mainstream’… But, progress has always be made because, one day, one guy stands up and say: “Wait a minute, is this really true?”
I agree with all that has been said on problems linked with missing values, outliers, good performance indicators, overfitting and underfitting, imbalanced classes, model validation, variable selection, leak variable detection, variable encoding, curse of dimensionality, deviation detection when applying a model, and descriptive power (ouch… This list is a good start for your other blog topic).
But, I do not agree when people (experts) are claiming that all these topics cannot be solved with automated processes providing good solutions in 95% of the cases, because that is what we (KXEN) have done (for each of the topics mentionned above). And the solution is very simple: even the experts use books, articles, and techniques that they have been trained on, I do not see any reason why a software could not use the same techniques in an automated way…
The real question is: ‘Is data mining technology mature enough in 2006 to solve automatically 95% of the business situations?’ My answer is clearly yes.
And this is linked with: where is data mining used today? I have read the example of the meteorologist, and I was thinking, for each meteo mathematical model, there must be 1 million models developped on Earth to detect “what customers will buy next month?”, “will my customer leave in the next 3 months?”, “how many customers will not reimbourse their credit”, “is this credit card transaction fraudulent?”, “how many products will I sell next week?”. If you count the number of mathematical models produced per year on our small planet, the very vast majority are in: CRM, Risk, Quality, and any kind of forecasts/predictions. There is a growing concern in the bio sphere, but, in terms of masses, it is nothing compared with CRM (a single KXEN customer generates thousands of mathematical models per year). For this, see http://www.kdnuggets.com/polls/2005/successful_data_mining_applications.htm
So the sentence of the book relates to the fact that, yes, we can fully automate Marketing Campaign Optimization; We can fully automate computation of Credit Risk probability of defect. Wherever we can translate a business problem in a suite of data mining tasks, we can automate now, today… with results comparable to a very good expert.

Be careful, nobody claims that, if you put very bright people for one year on the same problem, they will not get better results. Of course they will! But there is not enough experts on Earth for all the problems that can be optimized today using state of the art automated data mining techniques…

Now, this said, there are phases which will always be ‘human powered’ (I am waiting for a guy to stand up and say: “Wait a minute…”):
1/ How to go from a (business) problem definition to a decomposition of data mining tasks using data mining functions (I am NOT talking about algorithms here, but functions as defined in JDM).
2/ How to perform a business validation of the findings of the maths powered engines. This is why all useful data mining function implementations must be verbose and ’speak’ to business users in a way they can understand it.
3/ How to access to the data in a way a normal person (and not a database adminstrator) would like to. This is where collaboration with BI solutions is very interesting.
4/ How to discover new algorithms and techniques that will work in 97% of the cases…

OK. I am finihsed. I hope that I have managed to convey the passion of the other side of the force…

P.S. By the way, besides this one sentence, how was the book?? ;-)

31 Mapping Facebook’s Global Takeover

Facebook seems to be a bit of an omnipresent entity these days. Once the domain of college kids, it now features users of all ages and has also turned into a promotional tool for artists, politicians and businesses.

Vincenzo Cosenza has been tracking various social networks’ popularity throughout the world, and he just posted a December 2010 map — based on data by Alexa and Google Trends for Websites — showing just how prominent Facebook has become throughout the world.

A comparison between this new map and one from June 2009 reveals that Facebook — consistently the top social network in countries like the U.S., Canada and Australia is growing even more popular in the international sphere. Based on Cosenza’s findings, it looks like Facebook has replaced Orkut, once in the top spot, as the number one social network in India. And the site has also gained ground in South America and Europe.

There are some instances where other social networks have maintained their popularity; Russia is still dominated by V. Kontakte, and Orkut maintains a strong foothold in Brazil. Meanwhile, QZone’s big in China.

According to Consenza’s findings, other social networks, while not in the top spot, are also starting to gain some traction in other countries. LinkedIn is in third place in Australia and Canada, and Twitter holds the same ranking in Germany and Italy.

To get a better sense of how Facebook’s global usage has grown, take a look at the December 2010 and June 2009 maps below. The full graphic (including results from December 2009 and June 2010) may be viewed here.

World Map of Social Networking, Dec. 2010

World Map of Social Networking, June 2009

Via | Mashable.com

Tropiezos frecuentes en Data Mining

Data Mining a pesar de ser una disciplina "científica" posee mucho de artesanal. Un proyecto de modelización descriptiva o predictiva rara vez tiene un fin cierto y definitivo. Más bien, está constituido por una larga serie de ciclos de aproximación sucesiva que abunda en reformulaciones, prueba y error, intuición y especialmente, estrategia. El pensamiento estratégico es fundamental para lograr encontrar un buen camino, sin demasiados tropiezos, hacia la satisfacción del objetivo planteado. La ruta está llena de potenciales tropiezos que conviene evitar. Vamos a describir los más frecuentes.

Pocos datos

Es irónico que en una época de enorme cantidad de datos disponibles y donde uno de los principales argumentos de venta del data mining ha sido justamente la existencia de cuantiosos datos, uno de los tropiezos más frecuentes sea el de disponer de demasiado pocos datos. Sin embargo, este hecho es fácil de explicar. Las propiedades o sucesos interesantes para modelizar son justamente los muy poco frecuentes. Es poco útil modelizar (con una intención predictiva) sucesos que ocurren habitualmente. Si la propensión a la compra de un producto o servicio fuera suficientemente generalizada, y no un evento infrecuente, ¿para qué hacer un modelo o una campaña de marketing? Si el fraude fuera un suceso frecuente, no quedarían instituciones financieras y otras capaces de hacer modelos, habrían quebrado antes… Por lo tanto, la mayoría de los sucesos que vale la pena modelizar son infrecuentes, en muchos casos, muy infrecuentes. De allí el tropiezo frecuente de contar con demasiados pocos datos. Es importante que el analista se asegure antes de iniciar un proyecto de que cuenta con una cantidad apropiada de datos. Es inútil iniciar un proyecto con una cantidad insuficiente de datos.

Una sola técnica

Ser "especialista" en una sola técnica tiene algunas ventajas. Métodos como el análisis de regresión, redes neuronales o incluso árboles de decisión son muy ricos en parámetros, características, modalidades de uso, etc. Se requiere un conocimiento y experiencia considerables para su utilización experta. Son también métodos sumamente "potentes" en un sentido de capacidad modelizadora. En el caso de las redes neuronales, por ejemplo, existen teoremas que, bajo ciertas condiciones, afirman la "universalidad" de estas técnicas. Es comprensible de que haya analistas que trabajen utilizando solamente una de estas técnicas. Sin embargo, en la práctica esto puede ser limitante. Diversos estudios teóricos y prácticos no han sido capaces de producir conclusiones claras respecto de cuando una u otra técnica es apropiada (para qué tipo de problemas). Cada método ofrece perspectivas distintas e, incluso, complementarias sobre un problema y conjunto de datos. De modo que, en la práctica, emplear varias técnicas distintas enriquece un análisis y le permite al analista considerar un problema desde varios puntos de vista.

Sobreajuste

En el afán de obtener el modelo más preciso, más de una vez ocurre que el analista produce un modelo que "sobreajusta", es decir, un modelo bastante o muy preciso respecto del conjunto de entrenamiento usado, pero con una performance bastante peor sobre un conjunto de datos distinto, no utilizado para ajustar el modelo. Esto ocurre cuando se produjo o seleccionó un modelo que además de modelizar el suceso de interés incluyó también características idiosincráticas de la muestra de entrenamiento o "ruido" de esta muestra. Esto no solo no mejora un modelo sino que deteriora su capacidad predictiva cuando se aplica a otros conjuntos de datos pertenecientes a la misma población. Para evitar esta situación es fundamental la utilización de una muestra de validación, pero también de una medida de performance apropiada (en función del objetivo de la modelización y las propiedades de la variable target) y una interpretación correcta y experta de los diversos estadísticos y gráficos relevantes.

Sobrevalorar el "modelo campeón"

El "paradigma" habitual de modelización en data mining consiste en ajustar varios modelos a una muestra de entrenamiento y seleccionar aquel que posea la mejor performance predictiva. El problema es que aunque la medición se realice sobre una muestra de validación, esto no garantiza en absoluto obtener un modelo bueno o apropiado. Es necesario evaluar e interpretar un modelo a fin de determinar si es apropiado para el problema que se intenta resolver. Si no lo es, no hay modelo campeón que valga. Cuando esto ocurre, es fundamental diagnosticar la situación (muchas veces se trata de datos insuficientes, un problema mal planteado o algún otro tropiezo como los que se mencionan aquí) y volver a empezar.

Datos contaminados

¿Cómo se contaminan los datos? Utilizando variables que no estarán disponibles cuando quiera utilizar un modelo en producción. Estas variables tienen una correlación casi perfecta o directamente perfecta con la variable target (p. ej. la fecha de adquisición de un producto cuya propensión se quiere predecir). Muchas veces puede resultar obvio qué variables no debieran incluirse en una modelización, pero no siempre. Una señal de alarma es un modelo simple en el que una variable predictora genera una predicción perfecta. Durante las tareas de preparación y exploración de datos se debe estar especialmente atento a esta cuestión. Trabajar con datos contaminados es una pérdida de tiempo.

Datos atípicos

Los datos atípicos pueden tener un efecto nefasto sobre un modelo. Algunas técnicas de regresión, como las lineales, son particularmente sensibles a este problema. Durante la preparación y exploración de datos es importante estudiar y detectar el problema. También debe tenerse presente que "atípico" no es sinónimo de valor "extremo". Los valores extremos lo son en un sentido unidimensional, algo que es bastante más fácil de detectar. Los valores atípicos suelen serlo multidimensionalmente, algo que exige la utilización de técnicas estadísticas relativamente complejas para su detección (componentes principales, clustering, etc.).

Problemas heterogéneos

Un problema heterogéneo consiste en generar un modelo para un suceso de interés usando una población que si bien puede ser homogénea en algunos aspectos, no lo es respecto del suceso que se busca modelizar. Distintos segmentos de la población pueden poseer características lo suficientemente distintas como para hacer que las variables relevantes para una predicción sean diferentes o sus interacciones entre sí o con la variable target cambien cualitativamente. Hay que estar atento a las señales que marcan la posibilidad de la existencia de un problema heterogéneo porque sino, inevitablemente, la performance predictiva del modelo que obtendremos estará bastante por debajo de lo que hubiéramos podido lograr desarrollando modelos separados para cada segmento.

Subvaluar el conocimiento de dominio

Muchas veces, el analista de data mining no es un especialista del dominio específico en el que desarrolla sus modelos. Al menos no lo será durante cierto tiempo. Además, es probable que tenga una preferencia por hacer que "los datos hablen", en un sentido cuantitativo o estadístico, subvalorando conocimientos específicos de dominio que muchas veces no posee. Lamentablemente, los datos hablan en la forma de un "oráculo", sus mensajes requieren interpretaciones que muchas veces son subjetivas o diversas. Para comprender el mensaje correcto es fundamental complementar un conocimiento técnico solvente con un buen conocimiento de dominio que un experto seguramente podrá aportar.

Vía | SAS

11 Crucial Consumer Trends for 2011


RANDOM ACTS OF KINDNESS

When it comes to the mega trend of GENERATION G (that's G for Generosity, not Greed), there’s no better way for a brand in 2011 to put its money where its mouth (or heart) is than engaging in Random Acts of Kindness (R.A.K.). Consumers' cravings for realness, for the human touch, ensure that everything from brands randomly picking up the tab to sending a surprise gift will be one of the most effective ways to connect with (potential) customers in 2011 - especially beleaguered consumers in North America, Europe and Japan.

For brands, a serious (and sincere) R.A.K. strategy may mean no longer being seen as inflexible and unwieldy, but as more compassionate and charismatic instead. Something which is, of course, priceless and actually fun.

Fueling the R.A.K. trend is brands’ ability to actually know what’s happening in consumers’ lives (good or bad!), as people publicly and knowingly disclose (from Facebook to Twitter) more and more about their daily lives, their moods or their whereabouts.

Social networks also enable acts of kindness to spread far beyond its recipients, as they will gladly tell their friends and followers about the unexpected good news (see SOCIAL-LITES). Two fun examples to copy or improve on in 2011:

  • Flower delivery service Interflora has launched a social media campaign in the UK designed to brighten up the lives of Twitter users by sending them flowers. As part of the campaign, Interflora monitors Twitter looking for users that it believes might need cheering up. Once found, the users are contacted by tweet, and sent a bouquet of flowers as a surprise.
  • Dutch airline KLM’s ‘How Happiness Spreads’ Foursquare-based campaign employed a ‘Surprise Team’ to give passengers tailored, unexpected gifts at the airport. Throughout November 2010, as soon as someone checked-in at a KLM Foursquare location within its network of airports, the Surprise Team went online to find more background information about the person, decided upon a suitable gift and gave it them before they flew. For instance, one traveler tweeted he would miss a PSV Eindhoven football game while he was in New York. The Surprise Team, accordingly, gave him a Lonely Planet guide book of NYC with all the football bars highlighted in blue.

URBANOMICS

Urbanization remains one of the absolute mega trends for the coming decade. Here's just one telling stat: “Today, half the world’s population – 3 billion people – lives in urban areas. Close to 180,000 people move into cities daily, adding roughly 60 million new urban dwellers each year.” (Source: Intuit, October 2010)

How will this change the consumer arena in 2011 and beyond? Firstly, urban consumers tend to be more daring, more liberal, more tolerant, more experienced, more prone to trying out new products and services. In emerging markets, these effects tend to be even more pronounced, with new arrivals finding themselves distanced from traditional social and familial structures, while constantly exposed to a wider range of alternatives.

Secondly, keep a close eye out for ‘URBAN ISLANDS’: just 100 cities currently account for 30% of the world's economy, and almost all its innovation. Many are world capitals that have evolved and adapted through centuries of dominance: London, New York, Paris, etc. New York City's economy alone is larger than 46 of sub-Saharan Africa's economies combined. Hong Kong receives more tourists annually than all of India (Source: Foreign Policy, August 2010). However, metropolises such as Shanghai, Sao Paulo and Istanbul are obviously keen to join the top ranks, too.

Catering to city-citizens in these vast urban entities requires a local, dedicated approach in products, services and campaigns that mirror if not surpass the usual country-specific approach.

To cut a long story short: In 2011, go for products, services, experiences or campaigns that tailor to the very specific (and often more refined, more experienced) needs of urbanites worldwide, if not city by city. And don't forget to infuse them with a heavy dose of ‘URBAN PRIDE’. From Smirnoff’s Absolut Cities to BMW's http://www.bmwgroup.com/bmwgroup_prod/e/0_0_www_bmwgroup_com/home/home.html" target="_blank">Megacity vehicle, urban is the way to go.

Note: As URBANOMICS is perhaps one of the biggest yet still underestimated consumer trends of our times, we’ll expand on this trend in our February 2011 Trend Briefing, so if you haven’t done so already, please sign up for our free Trend Briefings >>>


PRICING PANDEMONIUM

While consumers have always looked out for special offers and discounts, new technologies and services mean that 2011 will see total PRICE PANDEMONIUM: 

  • More consumers are constantly connected, and when they hear about new deals online can quickly and easily spread them through their social networks.
  • Increasingly, consumers will be part of exclusive networks or groups to either receive special deals or demand them.
  • Mobile devices increasingly enable consumers to find or receive dynamic deals right at the point of sale, or to compare prices online. Case in point: Amazon.com just released an iPhone app that allows users to compare prices by scanning the product's barcode, photographing it or saying its name.

Always-on connectivity is changing consumer spending habits in myriad ways. For example, coupon clipping required planning and dedication, hence wasn't that popular with consumers more interested in the here and now (see NOWISM), but now is a near-effortless online activity. Furthermore, whipping out one’s smartphone at the counter, getting the latest deal via GPS, or barcode scanning is well, smart. And therefore a source of status rather than shame.

Brands will continue to respond with a host of innovative new business models and pricing strategies in the next 12 months, building on:

  • Group buying. The two billion consumers now online can exercise their collective buying power, helped by the host of services and social networks that make it easier than ever to organize and act. Keep an eye on 2010's big success story Groupon, which will become even bigger in 2011 (especially when part of Google ;-) or their competitor Living Social. Indeed, group buying sites are springing up everywhere, from GoNabit (that covers Arab markets such as Kuwait and Dubai), to Big Lion (Russia) and Daily Deal (Germany). Or consider these two examples as confirmation that group buying is here to stay: in September 2010, Chinese group buying site Taobao sold 200 Smart cars in 3 1/2 hours, while in October 2010 Walmart used Facebook to run their own group buying offer, which got the 5,000 'likes' needed to make the deal happen within 24 hours.
  • Member sales. The old 'club' format (think of Costco) has been given a new lease on life online, where niche communities thrive. Making some memberships limited or invitation-only, only increases the perceived exclusivity; SOCIAL-LITE consumers enjoy the social aspects of shopping, while for brands, offering reduced prices privately to small groups confounds TRANSPARENCY TRIUMPH. Designer fashion brands were some of the first to offer heavily discounted ranges to select groups, via sites such as vente-privee.com, Gilt Groupe and iDeeli but member sales are now expanding to areas such as travel (Jetsetter) and home furnishings (One Kings Lane).
  • Flash sales. Both groups and member communities frequently use time-limited offers that encourage impulse buys. By limiting the time available, and frequently only making sales available to members, brands are able to shift excess inventory quickly. DellOutlet and Threadless have seen great success with Twitter flash sales, and in November 2010 flash sale site Hautelook integrated its offers into its Facebook page, so that consumers could take up the day's deal without leaving the site. Indeed, with so many daily deals, there are now sites like Yipit and MyNines that aggregate all the deals.
    Less of a flash sale and more of a weekend-sale is retailer J. Crew's online factory store. Open every weekend from midday Friday to midnight Sunday (EST), the site offers a limited selection of some of J. Crew's most popular pieces, produced exclusively for the factory, at reduced prices.

  • Local discounts. With more and more consumers being able to broadcast their location, either publicly via Facebook, Twitter or other dedicated location-based services, brands can offer deals directly to consumers virtually at the point of sale. These can be rewards for performing certain actions (Shopkick and Checkpoints), geo-located promotions (PlaceCast), or just geo-enabled apps such as UK-based Vouchercloud. Indeed, despite the hype around check-in game services such as Foursquare, B2C brands are finding that the best way to incentivize customers is to offer them deals (witness Gap's promotion to launch Facebook Deals).
  • Dynamic pricing. Traditionally practiced by the airline industry, improvements in real-time information are now allowing other sectors to experiment with innovative dynamic pricing models, such as the US-based Off and Away, which auctions hotel rooms, and Swoopo, the German based 'entertainment shopping' site where every bid placed extends the auction's time period.

MADE FOR CHINA (IF NOT BRIC)

In 2011, expect an increasing number of 'Western' brands to launch new products or even new brands dedicated (if not paying proper respect) to consumers in emerging markets. After all, it's where the money is right now, and Western brands are still favored over local ones, so the combination of perceived quality with a bit of local tailoring, love or exclusivity makes total sense*.

* Like all consumers, these Chinese, Indian or Brazilian consumers will appreciate products that are tailored to their needs, wants and desires, either for practical reasons (shape, size, features) or because of the deep-rooted desire for recognition (cultural pride, heritage, lifestyles)
.

This MADE FOR CHINA (IF NOT BRIC) phenomenon is just one of the many sub-trends spawned by the macro trend of economic and consumption power shifting towards emerging markets. Just to remind you:

  • China's retail sales, the main gauge of consumer spending, rose 18.7% year on year to USD 183 billion in May 2010, following a 15.2% rise the previous year. (Source: Chinese National Bureau of Statistics, June 2010)
  • China’s retail sales may outstrip those of the US by reaching USD 5 trillion in 2016. (Source: Chinese People’s Political Consultative Conference, September 2010)
  • Affluent Chinese consumers prefer foreign brands: 52% of consumers whose annual income exceeds RMB 250,000 (USD 36,765) trust foreign brands more than Chinese ones while just 37% said they prefer the latter. (Source: McKinsey, September 2010)

So who has already jumped on the MADE FOR CHINA band wagon? Check out the following examples:

  • Levi's dENIZEN Jeans brand, targeting Asians/Chinese consumers with slimmer fits.
  • Dior's very expensive Shanghai Blue Phone, only available in Shanghai stores.
  • Hermès’ new Chinese brand, Shang Xia; its luxury stores sell ready-to-wear and decorative arts inspired by Chinese culture.
  • Chloé’s limited edition Marcie handbag to celebrate their fifth anniversary in China.
  • BMW’s limited edition, orange metallic M3 Tiger to celebrate the 25th anniversary of its M3 model in China, coinciding with the Chinese lunar calendar (the year of the tiger).
  • Last but not least, and the inspiration behind the naming of this trend: Apple's Shanghai store employees started wearing red t-shirts with the slogan 'Designed in California, Made for China' written in Mandarin. The message is a play on the words that are found on the back of all iPhones: 'Designed by Apple in California, assembled in China'.

P.S. Brands like Honda (Li Nian), Nissan (Venucia) and GM (Baojun) are busy introducing ‘cheaper’ brands in China, too, as many consumers in emerging markets still do have less to spend than their counterparts in mature consumer markets. And yes, that's yet another 'Emerging Markets' sub-trend. If you really can’t get enough of this topic, do re-read our Trend Briefings on FUNCTIONALL and EXCEPTIONALL.


ONLINE STATUS SYMBOLS

Online culture still is the culture, and thus we'll see a rise in online status symbols in 2011 (after all, status symbols reflect the zeitgeist like nothing else). What started with showing off the number of visitors to one's Flickr pages or blog now also encompasses the number of one's Facebook friends (or any other social network), Twitter followers, Foursquare check-ins and a host of other metrics that indicate one's 'wiredness'.

In 2011, you can’t go wrong supplying your (online-loving) customers with any kind of symbol, virtual or 'real world' that helps them display to peers their online contributions, interestingness, creations or popularity.

Indeed, one extra element to watch out for in 2011 is new status symbols that straddle the 'real' and 'online' worlds. From physical manifestations of digital status (think personalized Facebook and Twitter memorabilia), to online recognition of physical activities (status updates or badges based on real-world visits), consumers will seek to display their online status symbols in all arenas. Some fun (yet telling) examples:

  • Twournal enables users of Twitter to transform their tweets and pictures into a real-life published journal. In addition to creating their own 'books', users can also buy and sell publications from other users.
  • US based CrowdedInk offers an app that allows users to generate mugs filled with pictures of their Facebook friends or Twitter followers. Users only need to enter their username and a preview of the mug is automatically generated in minutes.
  • Location-based social game Foursquare awards members with badges for performing tasks, including the Supermayor badge (awarded when someone is mayor of 10 different places at once), the Entourage badge (awarded when checking-in with 10 friends), the Gym Rat badge (awarded when someone has checked- in on 10 trips to the gym in 30 days), and even the Last Degree badge (awarded when checking-in at the North Pole ;-).
  • Nerd Merit Badges are real-life, physical representations of a user's online achievements through Foursquare. The fully embroidered, velcro-backed badges sell for approximately USD 6. And Foursquare's own online store now offers similar real-world buttons.

WELLTHY

As good health is now as important to some consumers as having the biggest, newest or shiniest status symbols, growing numbers of consumers will expect health products and services in 2011 to prevent misery if not improve their quality of life, rather than merely treating illnesses and ailments. Some signs of the times:

  • 73% of US consumers consider being physically fit important to being ‘well’, with 74% including ‘feeling good about themselves’. (Source: The Hartman Group, August 2010)
  • An estimated 500 million people worldwide are expected to be using mobile healthcare applications by 2015. (Source: Reasearch2Guidance, November 2010)
  • There were nearly 17,000 health apps available in major app stores in November 2010, with 57% of them being aimed at consumers rather than health care professionals. (Source: Reasearch2Guidance, November 2010)
  • The heaviest use of health or medical related apps is by young adults: about 15% of those aged 18 to 29 have such apps, compared to 8% of users aged 30 to 49. (Source: The Pew Internet Project, October 2010)

In 2011, count on even more monitoring technologies becoming portable or even wearable, as well as getting cheaper (the smartphones held by many consumers are now more advanced than most dedicated medical devices).

Also, both regular and dedicated medical social networks give audiences a platform to share, compare and discuss their personal health issues with other consumers.

Last but not least, the 'consumerization' of health means that more consumers will choose products with embedded health benefits that are actually well designed, desirable, accessible, fun, tasty, interesting or storied. Some examples:

  • The Strollometer is a device that tracks all aspects of a new mother's strolling routine. On the display screen of the gadget, moms can check their speed, distance traveled, time spent exercising, average and maximum speeds, and the temperature outside. Once the strolling session is over, moms can then enter their data on the Fit4Mum.com website and see their results in terms of calories burned.
  • Sleep On It is a mobile app that allows users to track their nightly sleeping patterns. The free tool can determine what factors may be affecting their sleep and how much sleep they need each night to feel rested and energized. Sleep On It allows users to track the length and quality of their sleep, as well as time spent snoozing and mood to see how sleep impacts their overall health and quality of life.
  • Phillips DirectLife is a small, lightweight, wearable monitor that builds up a detailed record of users' daily activities. Users' data is uploaded to their personal DirectLife site, showing how many calories were burned throughout the day. Users track their progress against their personal targets, and have access to an online personal coach. Users can also compare themselves to other members and connect with the online community to encourage them to share tips and stay motivated.
  • Both Microsoft and Sony have recently released motion-sensing game controllers: the X-Box Kinect and Playstation Move. The systems use cameras (and a controller for the Move) to detect users' movements during gameplay. The controllers were designed not only to heighten the gaming experience but also, much like the Nintendo Wii, to add a physical and healthy dimension to it, with games like EA Sports Active 2 and Get Fit with Mel B.


SOCIAL-LITES AND TWINSUMERS

In 2011, word of mouth and recommendations will be even more dependent on P2P dynamics.
If TWINSUMERS* (consumers with similar consumer patterns, likes and dislikes, and who are hence valuable sources for recommendations on what to buy and experience) are all about improving 'search curation', SOCIAL-LITES are all about discovery, as consumers become curators; actively broadcasting, remixing, compiling, commenting, sharing and recommending content, products, purchases, experiences to both their friends and wider audiences.

* An update on our TWINSUMER trend will find its way to you in Q1 of 2011. That is, if you're signed up to receive our monthly Trend Briefings.

Why would consumers want to become curators? Because many of them are investing time and effort in building BRAND ME, via online profiles that record their opinions and recommendations. And as audiences in knowledge economies value interesting, relevant and useful tidbits, they bestow status on those curators or SOCIAL-LITES who share.

Furthermore, it's never been easier to be a SOCIAL-LITE. Rather than having to call, text or even email people personally, social networks' streams allow users to easily broadcast information to a wide range of people without interrupting or intruding.

So, consumers will talk more about brands in 2011 than ever before, and opportunities for brands that create engaging content that consumers want to share, or that have personalities that actually engage consumers will also be bigger than ever. Making it easy for SOCIAL-LITES to retweet or 'like' this content is of course requirement number one. Examples:

  • The STRP Art and Technology Festival which ran from 18 - 28 November 2010 in Eindhoven, the Netherlands, employed the use of RFID in museums and galleries that allow guests to rate art via RFID wristbands. Visitors used RFID wristbands and 'dashes' (which represent how much they liked something) to tag art, with the results collated to create a visible "tag cloud," showing the most popular exhibits, and where they were located.
  • Gogobot is an online travel community with a social lens. Users ask questions about destinations (think 'Where's a fun restaurant in Paris?'), not just to the Gogobot community but also to their Facebook and Twitter networks. Gogobot collates the answers, and includes pictures and links to all of the places mentioned.
  • Levi's has integrated its online store with Facebook, allowing shoppers to socially interact with friends and create a 'like minded shopping' experience. From the Levi's store, users are able to share products with friends through the Friend Store, 'Like' them or search for 'Top Liked' products within Facebook.
  • Amazon now enables users to integrate their Facebook and Amazon accounts. The feature allows Amazon to connect through to a user's social network, then base recommendations (think books, DVDs and musicians) upon the information found in his/her Facebook profile.
  • US based Mombo analyzes Twitter feeds from users all over the world and assigns ratings on movies based on the collective opinion of Twitter users, with each tweet being analyzed by the site's 'sentiment analysis' engine. Users who sign in with Twitter will get movie recommendations based on what they and their friends have tweeted about.
  • Storify helps its users to curate content that they find on social networking sites. Users cut and paste pieces of content (including pictures and video) from various sources on the web and put it all together in order to communicate their desired message.

Warning: Social customs and behaviors are still evolving. But the connections that consumers establish with one another are personal and therefore valuable (and not all about brands!). And SOCIAL-LITES still operate in a social sphere with intrinsic importance to participants. As such, brands need to tread carefully, treating consumers with respect, and be completely transparent. Get it wrong and you'll find 2011's SOCIAL-LITES can be bitchy as hell ;-)


EMERGING GENEROSITY

In 2011, GENERATION G(ENEROSITY) will continue to give. Consider EMERGING GENEROSITY, which is about brands and wealthy individuals from emerging markets (yes, especially China) who will increasingly be expected to give, donate, care and sympathize versus just sell and take. And not just in their home countries, but on a global scale. It's a profound cultural change and a consumer demand that their counterparts in mature markets have had a few years to getting used to. Some fun stats:

  • 86% of global consumers believe that business needs to place at least equal weight on society’s interests as on business’ interests.

    78% of Indian, 77% of Chinese and 80% of Brazilian consumers prefer brands that support good causes, compared to 62% of global consumers.

    8 in 10 consumers in the India, China, Mexico and Brazil expect brands to donate a portion of their profits to support a good cause. (Source: Edelman, November 2010.)

  • The number of millionaires in India in 2009 grew 51 percent, to 126,700. (Source: Merrill Lynch, June 2010.)
  • 88-year old Yu Pengnian became China’s first billion dollar philanthropist in April 2010. The Yu Pengnian Foundation today has USD 260 million of bank deposits, and a Hong Kong and Shenzhen property portfolio worth just under USD 1 billion, which is expected to contribute USD 50 million of cash each year to the foundation. (Source: Hurun Rich List, October 2010.)
  • In September 2010, Bill Gates and Warren Buffett hosted a dinner for China's ultra-rich to promote philanthropy. The event was similar to their efforts to get US billionaires to pledge to give away at least half their wealth, although the two denied that they were recreating the pledge or pressurizing people to give, merely using the event to answer questions and discuss issues around developing a culture of philanthropy. (Source: Washington Post, October 2010.)

In 2011, any brand or individual doing exceptionally well, will be expected to join GENERATION G. Thanks to shifts in power and a relentless TRANSPARENCY TRIUMPH, the ones who don't know how to give will have no excuses left, and nowhere to hide ;-)


PLANNED SPONTANEITY

Two years ago, we highlighted NOWISM: consumers' lust for instant gratification being fuelled by ever quicker, more accessible real-time products and services.

We also noticed that NOWISM was part of a larger cultural, social and technological trend towards greater spontaneity. With lifestyles having become fragmented, with dense urban environments (URBANOMICS!) offering consumers any number of instantly available options, and with cell /smartphones having created a generation who have little experience of making (or sticking to) rigid plans, 2011 will see full-on PLANNED SPONTANEITY.

Expect to see consumers in 2011 rushing to sign up to services (the PLANNED part) that allow for endless and almost effortless MASS MINGLING with friends, family, colleagues or strangers-who-may-become-friends-or-dates (the SPONTANEITY part ;-)

What's next? Look beyond Twitter, and expect (younger) consumers voluntarily opting in to passively and continuously share their location*, in return for truly smart suggestions about what they could be doing or who they could be meeting up with. Will consumers accept their mobile data being collected, aggregated, analyzed and shared? If the benefits are interesting, fun or life-enhancing enough, then most likely yes.

*According to Jeff Jonas at IBM, mobile devices in the US are creating approximately 600 billion geo-spatially tagged 'transactions' a day (that's every call and text, not just web connected smartphones!)

For consumers, knowing where they are and what's / who's around them is the key to PLANNED SPONTANEITY. That's about to get a whole lot easier, as geo-location becomes a key feature of social networks and web apps (from existing providers adding location information, such as Facebook's Places, Twitter's locator, and Google's Hotpot, to dedicated services like Foursquare, Gowalla and Brightkite). Examples:

  • Geomium takes data from local review sites like Yelp and Qype, and combines it with social information to not only allow users to both see which of their friends are nearby, but also to find nearby event and venue information and deals.
  • LikeOurselves lets individuals with shared interests find each other via mobile groups. The service also allows users to quickly create a group and locate members within 20 miles of their location, enabling on the fly meetups.
  • Fast Society is a New York-based startup that has returned to Twitter's roots, aimed at simplifying communication between friends on-the-go. The service is SMS-based, and groups last between 3 hours and 3 days, increasing the spontaneous nature of the offering.
  • Unsocial aims to be a facilitator for people to meet others that share the same profession or industry. Unsocial works using a location-based algorithm - by opening the app and logging in a user is able to press the 'People' button, after which the app will display relevant matches nearby.

ECO SUPERIOR

The number one challenge for governments, consumers and businesses (recession or no recession) in 2011 remains the quest for more environmentally sustainable societies and economies. When it comes to 'green consumption', expect a rise in ECO-SUPERIOR products: products that are not only eco-friendly, but superior to polluting incumbents in every possible way. Think a combination of eco-friendly yet superior functionality, superior design, and/or superior savings.

Why the need for ECO-SUPERIOR? The number of consumers actively seeking out 'green' products is reaching a plateau, as mainstream consumers start to question the value and efficiency of going green:

  • While 40% of consumers say they are willing to purchase green products, only 4% of consumers actually do when given the choice. (Source: Journal of Marketing, September 2010)
  • 58% of global consumers think that environmentally friendly products are too expensive, while 33% of global consumers think that environmentally friendly products don't work as well. (Source: GfK Roper, September 2010)
  • While the volume of green products available to US consumers increased by 73% between 2009 and 2010, only 5% of products were not found to include some 'greenwashing' claims. (Source: Terrachoice, October 2010)

Expect to see a number of leading brands in 2011 switch from purely marketing their products' sustainability and eco-friendliness (with its niche reach) and taking aim right at the heart of traditional alternatives: stressing the superior quality and design, increased durability and/or lower running costs of products in ways that will appeal to even the most eco-skeptic, self-centered or financially-challenged consumer.

ECO-SUPERIOR examples? How about the Ovopur, a well-designed natural and environmentally friendly water purification device, or the Stealth Toilet, which contains a flushing system that only uses 0.8 gallons of water per flush - saving the average family approximately 20,000 gallons of water each year, or the Renault DeZir; a 'green' concept supercar that travels from 0-60 in five seconds, or Philips' 12-watt EnduraLED bulb...

Meanwhile, our ECO-EASY trend, which we highlighted in our 10 trends for 2010, is still going strong. This is what we said a year ago:

"While the current good intentions of corporations and consumers are helpful, serious eco-results will depend on making products and processes more sustainable without consumers even noticing it, and, if necessary, not leaving much room for consumers and companies to opt for less sustainable alternatives to begin with. Which will often mean forceful, if not painful, government intervention, or some serious corporate guts, or brilliantly smart design and thinking, if not all of those combined. Think anything from thoroughly green buildings, to a complete ban on plastic bags and bottles, to super-strict bluefin tuna quota — anything that by default leaves no choice, no room for complacency, and thus makes it 'easy' for consumers (and corporations) to do the right and necessary thing".

Here's just one ECO-EASY example that recently caught our eye:

  • In September 2010, Italy's famous Cinque Terre national park announced that the use of plastic water bottles by tourists and visitors was forbidden in an attempt to preserve the coastline from pollution and litter. Instead, visitors will be able to purchase reusable metal flasks which they will be able to fill up with still or sparkling water from public fountains.

OWNER-LESS

We covered the re-emergence of fractional ownership and lifestyle leasing business models (no more dodgy timeshares!) in our TRANSUMERS briefing way back in 2006. Brands like Rent the Runway (fashion) and Avelle (handbags - formerly Bag, Borrow or Steal) and P1(luxury cars) have shown that for many consumers, access is better than ownership. Indeed, over the past few years, there have been few industries that haven't got the 'Netflix treatment', from textbooks to jewelry to educational video games to calculators.

For consumers, the appeal is obvious:

  • Traditional ownership implies a certain level of responsibility, cost and commitment. Consumers looking for convenience and collecting as many experience as possible want none of these things.
  • Fractional ownership and leasing lifestyle businesses offer the possibility of perpetual upgrades to the latest and greatest, the ability to maximize the number and variety of experiences, and allow consumers to access otherwise out-of-reach luxuries.
  • Owning bulky, irregularly used items is both expensive and unsustainable, especially in dense urban environments where space is at a premium. With more consumers having mobile access to online systems, it becomes easier to book items whenever and wherever they are needed (see PLANNED SPONTANEITY).

Now, 2011 could be the year when sharing and renting really tips into mainstream consumer consciousness. Two key developments:

  • In 2011, expect to see more and more big brands getting in on the action. Take for example car sharing, one of the great successes of the OWNER-LESS trend, with car clubs springing up all around the world: Zipcar is the market leader, but similar services can be found everywhere from Australia (GoGet) to Brazil (Zazcar). Big brands, having seen the success of these smaller startups are increasingly getting in on the action: Hertz launched their car sharing service Connect back in December 2008, and now Daimler has added Hamburg and Austin, Texas following the successful pilot of their car2go pilot in Ulm (Germany). In July 2010, Peugeot launched its Mu 'mobility' service in the UK after successful launches in France, Germany, Italy and Spain. Customers can rent cars, scooters, vans or even bicycles.

  • Local authorities are finding that shared solutions allow them to expand their services at a lower cost, and in a sustainable manner. Public bike programs have been the global hit of 2010, with schemes launching in Minnesota, London, Mexico City, and across the Ruhr in Germany. Now, governments are exploring new forms of transportation, with Paris (who pioneered bike sharing) launching Autolib, an electric car sharing scheme in September 2011, and the New York City Department of Transportation announcing in October 2010 that they had partnered with Zipcar.

Another big boost for the OWNER-LESS economy is that with so many highly visible transportation initiatives, all consumers are becoming used to seeing schemes in action, and more and more are feeling comfortable with the idea of sharing and renting large, expensive or often-idle objects.

P.S. The brief introduction above only looks at brand/ government offerings. There is also a whole host of P2P sharing sites that are springing up to facilitate renting and sharing between consumers, with everything from homes to fashion, to cars (and car parking spaces) available. Watch this space (Q1 of 2011) for a full Trend Briefing on the rise of OWNER-LESS.


12. NEED MORE?

For many of you, this Trend Briefing is probably enough to keep you going for a couple of months when looking into and working with consumer trends. And yet, the above is just a snapshot of what we track.

So, if you need access to all the trends we’re tracking for 2011 and beyond, including fully indexed and searchable trend examples from around the world, then please check out our Premium Membership service.

It includes password-protected access to a full, 100+ slide/page 2011 Trend Report and to our ever-growing Trend Database which now contains 4,000+ examples. Premium clients also receive monthly updates and bonus content until November 2011.


You’ll find yourself in good (yet pretty competitive) company: Leading B2C brands and agencies (see the sample above), from Google to Virgin to Saatchi & Saatchi, are already working with our Premium content. Many ambitious small firms have purchased too, as we’ve kept things pretty affordable. More info here »


Apply!

For loyal readers, this is by now old news: the four ways to apply these consumer trends, and make some money from the innovations they spawn. Just ask yourself if they have the potential to (and if so, how):

  1. Influence or shape your company's vision.
  2. Inspire you to come up with a new business concept, an entirely new venture, a new brand.
  3. Add a new product, service or experience for a certain customer segment.
  4. Speak the language of those consumers already 'living' a trend.

For more tips on how to spot and apply trends, see our TIPS section. Easy peasy, no? Have fun in 2011!

Via | TrendWatching

Geomarketing, transeúntes, más aproximaciones. City Analytics

En un post anterior comentaba cómo parecen abrirse nuevas metodologías de cálculo del número de personas que transitan por un determinado tramo de vía. En aquel caso, a través de los dispositivos electrónicos con GPS, que permiten la geolocalización de sus portadores.

City Analytics es un proyecto de la empresa Ciudad 2020, que se presenta como el Google Analytics del mundo real (¡nada menos!), aplicando los principios y perspectiva metodológica de Reality mining. Se basa en detectar los dispositivos bluetooth que pasan delante de un negocio, gracias a una antena instalada en un PC y el software correspondiente. Los datos monitorizados son enviados a los servidores de City Analytics, tratados y analizados, de manera que el usuario sólo tiene ya que conectarse para acceder a los informes de tránsito frente a su negocio, básicamente:

  • número de transeúntes, diferenciando peatones de vehículos, tiempo de parada,
  • por horario, día de la semana, festivos vs laborables, períodos vacacionales…
  • tipo de transeúntes: residentes, habituales, ocasionales; si repiten o no

Los cálculos para obtener usuarios únicos, discernir el medio de transporte y demás requieren evidentemente algunas estimaciones y asunciones sobre el comportamiento de estas personas. Por ello, las cifras absolutas arrojadas por el sistema hay que tomarlas como una aproximación. Y habrá que hablarlo con ellos a fondo para valorar con precisión cuál puede ser el error cometido -de momento sólo lo hemos hablado a nivel general-.

Por otro lado, no todos llevamos dispositivos con bluetooth activado, muchos llevamos más de uno… Sea como sea, siempre sobre una muestra cuasi-exhaustiva y con una toma de datos continua, todo un lujo. Y además, a un coste asumible por cualquier pequeño comerciante.

Ésta es una primera aplicación, muy interesante, para un comerciante o una enseña minorista de cualquier sector. La otra gran aplicación está en las inferencias y estimaciones de flujos de personas para ciudades completas. Esto será posible una vez tengan instalado el número suficiente de antenas y equipos tomando datos en una ciudad, y subiendo a sus  servidores estos datos. Y eso nos interesa mucho a quienes estamos continuamente buscando estimadores de flujo de transeúntes para aplicarlos a nuestros estudios de geomarketing.

Vía | unica 360

Automating Analytics

Information Management has recently published an interesting article about automating analytics. Written by Bill Franks and Scott VanValkenburgh, it explains when automating analytics work and when it doesn’t.

I think automating data mining is in the top five open issues in the field. It’s not an easy task, but it can be very time (and thus cost) saving when one need to generate hundreds or thausands of models each day. According to the authors (and I agree with them):

The reality is that there is no magic button. [...] The initial processes still need to be built and automated. Maintenance and modifications also need to be performed on the processes over time as business needs, data structures or other factors change.

I think some tasks such as ETL, scoring and quality checks can be automated while ressources will be free up for data miners to build models. If you’re interested in this subject, you may want to read an older (2006) post with a lot of interesting comments on Data Mining Research: Garbage in, Garbage out. I think that my mind has changed from four years ago. At least, I’m convinced that some parts of the data mining process can be automated. However, the limits of automating data mining are still not clear to me. I guess it’s an open issue for the data mining field.

One good point highlighted by the authors is the life duration of a model. They advised not to “try to use an existing model for a brand new question, and don’t let a process outlast its reasonable life span.

Via | DataMiningBlog

Una curiosa campaña de Toyota consigue que los hispanos luzcan en su coche el logotipo con orgullo

Para recuperar el terreno perdido tras una crisis motivada por una serie de fallos en algunas piezas, Toyota está realizando una acción en Estados Unidos dirigida a los hispanos que ha logrado que más de 282.000 personas luzcan en sus coches una curiosa pegatina regalada por la marca de automóviles.

Una curiosa campaña de Toyota consigue que los hispanos luzcan en su coche el logotipo con orgullo

Según Advertising Age, a la agencia Conill se le ocurrió la idea de regalar vía Facebook pegatinas en las que, además del mensaje “Somos muchos Toyota”, apareciera otro según el origen del usuario: “Somos muchos mexicanos”, “Somos muchos argentinos”, “Somos muchos puertorriqueños…”. Un total de 99 opciones, que incluyen gentilicios de países y de ciudades latinas.

La agencia también sacó sus cámaras a la calle para explicar desde un iPad la iniciativa a los hispanos propietarios de Toyota (fueron a grandes aparcamientos de estadios de fútbol o centros comerciales). Mientras los consumidores pedían a través de la tableta de Apple sus pegatinas y contaban su historia con la marca, eran grabados. Con este material se han creado 15 spots, que se han colgado en internet. Seis de ellos se están emitiendo además en televisión.

Hasta ahora más de 282.000 usuarios de Facebook han solicitado su pegatina y el perfil de Toyota Latino suma más de 25.000 seguidores. La pegatina más pedida es la de “Somos muchos mexicanos”, lo que refleja el hecho de que las dos terceras partes de los hispanos de Estados Unidos tiene su origen en el país vecino.

Según datos de la agencia PR Tracker, desde que la campaña diera comienzo, en julio, la opinión favorable de los consumidores hacia Toyota ha mejorado un 13% y la consideración por vehículos Toyota, un 8%.

Vía | MarketingNews