By mapping individual neurons in the brain with red, green and blue derivatives of fluorescent protein, Jeff Lichtman and Joshua Sanesinvented the art of ‘brainbows’ in 2007 at Harvard Medical School. This image shows a transgenic mouse hippocampus magnified by 40 times. Image courtesy of Livet, Weissman, Sanes and Lichtman, Harvard University.
Welcome to THINK QUARTERLY tHe DAtA ISSUe At Google, we often think that speed is the forgotten ‘killer application’ – the ingredient that can differentiate winners from the rest. We know that the faster we deliver results, the more useful people find our service. But in a world of accelerating change, we all need time to reflect.Think Quarterly is a breathing space in a busy world. It’s a place to take time out and consider what’s happening and why it matters. Our first issue is dedicated to Data – amongst a morass of information, how can you find the magic metrics that will help transform your business? We hope that you find inspiration, insights, and more, in Think Quarterly. Matt Brittin managing Director, UK & Ireland operations Google 05
ConTriBUTorS Simon rogerS Ulrike reinhard Sarah BraCkingThink Quarterly represents Simon Rogers is the editor Ulrike Reinhard is the editor Sarah Bracking is the authorthe imagination, insights of guardian.co.uk/data, which of WE magazine, a digital of Money and Power andand knowledge of a global encourages users to visualise publication focussed on Corruption and Development,community of contributors raw datasets. He has edited two emergent net culture. With and is a Senior Lecturer infrom journalists and academics Guardian books: How Slow Can Peter Kruse, she co-founded Politics and Development atto industry experts and Google you Waterski? and The Hutton the What’s Next? Institute the University of Manchester.insiders. They’re united by a Inquiry and Its Impact. In to research cultural value She is a member of thepassion for cutting-edge ideas 2010, Simon received a special preferences and social International Institute forand seeking out the points of commendation from the Royal megatrends. She interviews Democracy and Electoralfundamental change in the Statistical Society in its awards Hans Rosling on page 16; Assistance and has workednew era of digital business. for journalistic excellence. writes about the online with the Norwegian Agency He interviews Vodafone CEO video ad boom on page 34; for Development Cooperation. Guy Laurence on page 10, and speaks to Peter Kruse She writes about ethical and chooses his top 10 data about harnessing collective investment in the developing websites on page 28. intelligence on page 50. world on page 22. 06
ConTenTSContents pages can be bossy– telling you where to go andwhat to do. But they can beuseful, too. After all, what is acontents page but data aboutthe book you’re holding? Still,we thought we’d have a littlefun with it. You can peruse thisbook in the usual linear fashion,but look a little closer and newdata relationships emerge.Don’t just read: explore. PAGE 10 16 22 28 30 34 EXECUTIVE A DATA STATE DATA FOR THE LUNCH FULLY INSIGHT OF MIND CHANGE KNOWLEDGE WITH HAL VIRAL SIMON ROGERS ULRIKE REINHARD SARAH BRACKING SIMON ROGERS HOLLY FINN ULRIKE REINHARD OPEN DATA OPEN DATA OPEN DATA GRAPHICS GOOGLE CREATIVITY DATA OBESITY CREATIVITY DECISION-MAKING TOP 10 STATISTICS ADVERTISING GUY LAURENCE INTERACTION PRIVATE INVESTMENT GOVERNMENT BUZZMAN VODAFONE GAPMINDER FOUNDATION SOCIAL CHANGE DATA OBESITY FRENCH CONNECTION SOCIAL CHANGE GAPMINDER FOUNDATION EXPERIMENTS YOUTUBE GOVERNMENT GLOBAL DEVELOPMENT HAL VARIAN VIRAL MARKETING HANS ROSLING CORPORATE RESPONSIBILITY 08
38 40 44 50 56 60 62FROM STICKS AD OPEN FOR SOFT VALUES, FREE YOUR LONG FLIGHT THE MOBILE TO CLOUDS VALUE BUSINESS HARD FACTS POCKETS AHEAD REVOLUTION GRAPHICS TONY FAGAN NIGEL SHADBOLT ULRIKE REINHARD RICH PLEETH GRAPHICS GRAPHICS OPEN DATA GOOGLE OPEN DATA DECISION-MAKING GOOGLE PUZZLE OPEN DATA SOCIAL CHANGE STATISTICS DECISION-MAKING SOCIAL CHANGE SOCIAL CHANGE TRANSPORT ADVERTISING PRIVATE INVESTMENT PETER KRUSE TRANSPORT EXPERIMENTS GOVERNMENT NEXTPRACTICE INTERACTION COLLECTIVE INTUITION NEAR FIELD COMMUNICATION MOBILE INTERNET 09
ExEcutivE Guy Laurence, ceO Of VOdafOne uK, KnOws a thinG Or twO abOut infOrmatiOn OVerLOad. feeLinG stressed Out by statistics? he has the cure fOr data impOtence. wOrds by Simon RogeRS pOrtraits by SpenceR muRphya few seconds after midnight on new year’s have a relationship with numbers, i haveeve, 2010. numbers start flying across a a relationship with customers,” says Guybank of screens in a large darkened room. Laurence, the 49-year-old who took overLondon: 1,170,000; Glasgow: 115,000; as ceO of Vodafone uK in 2009. “i focusmanchester: 75,000; Leeds: 70,000… the totally on human responses to things;numbers scroll on as the black-clad tech if you smack someone in the face, whatteam look for signs that the system might would they do? if you kiss them on thenot be able to cope. cheek, what would they do?” it could be a scene from futurist cult Laurence took over a company widelyfilm Minority Report, but the room is seen as stagnating in third place in theactually a real one – at Vodafone’s state- uK’s competitive mobile market. today,of-the-art network Operations centre Vodafone is viewed as a powerful successin newbury, berkshire – and the figures story, with more than 19 million customersrepresent the number of texts sent in across the country. when you’ve got thatthe first 30 minutes of 2011. this is pure many customers, the big question is:data in action. how do you industrialise something that the man responsible for this scene is works for each one? “you can always kissobsessed with data – because of what the one customer on the cheek – but hownumbers can help him do, rather than with do you kiss 19 million customers on thethe ones and zeros themselves. “i don’t cheek?” he asks. iNSiGHt 11
Laurence carries only a few numbers in feel a network,” he says, “but at the nOc information that companies once believedhis head: his company’s ‘net promoter score’ we absolutely live and breathe data in real was commercially confidential is now(which tells him exactly how well Vodafone time.” managing 90 million calls and 80 routinely published – or leaked to websitesis really doing with its customers) and million texts on an average day is a tricky like wikileaks.the competition’s market revenue share. business; a typical 24 hours sees Vodafone Laurence says he is ‘relaxed’ about“when you run a £5 billion company you carry 45 terabytes of data, equivalent to increased demands for transparency.can’t avoid numbers – but if you start with 11.25 million music tracks. “companies will become more transparentnumbers you’ll never innovate,” he says. Vodafone’s approach is to use data to as a necessity – customers now see that“you have to take the action you think will manage demand before things happen. the as an essential part of the trust equation.”work and the numbers follow.” company’s plans for the royal wedding in the bigger impact may come from the even when he’s about to fly off with his april include adding extra temporary base technology that is making access tofamily to live rough in the masai mara for stations to cope with heavy network usage. this data a mobile phenomenon. “thisa week, for Laurence, it’s all about focus. he when take that tickets went on sale just industry is de-linking access to data fromleft school with one grade e a-level, having before christmas and the band’s official physical location,” he says. in a world wherefluffed his exams by setting up a candle- website crashed due to demand, Vodafone shoppers can check out the competition’smaking business after he realised that was prepared for the surge of fans texting prices while they’re in your store, keeping“making money was much more fun”. it’s one another to check whether they’d got control of data is no longer an option.a pattern repeated when he quit his degree their tickets. but for now, managing the informationto work for independent music publisher One of the walls at Vodafone’s out there is the priority. access tochrysalis. eventually he became head of operations centre shows connections to information was once the big problem, saysdistribution and marketing outside america 217 countries to monitor how much traffic Laurence. then it quickly flipped, throughat mGm. his job was to work out which is coming in from abroad in real time. technology, to data overload. “we weremarkets a product would work in. the data shows that different cultures are brought up to believe more data was good, he will tell you, for instance, that a ‘asymmetric’, says Laurence. “you can see and that’s no longer true,” he argues.baseball movie will only work outside the polish mothers are texting their sons over Laurence refuses to read reports fromus if it’s shown in Japan. he worked on the here to see if they’re okay, but the sons are his product managers with more than fivebond films, including GoldenEye, selling not texting back,” he says. “but the french of the vital key performance indicators onthem to reluctant cinema owners who hadn’t are almost symmetrical – so as the texts go them. “the amount of data is obscene. thescreened anything from the franchise in six out, the replies come back in. as situations managers that are going to be successfulyears. “the last film had been [classified as] unfold in real time in egypt or bahrain we are going to be the ones who are prepareda 15. therefore anyone under 21 had never can see how that affects the network, too.” to take a knife to the amount of data…seen a bond film in a cinema.” so mGm even a bill being sent by email triggers Otherwise, it’s like a virus.made it cool – selling the film to teenagers, a whole chain of data events: customer gets “where did it all go wrong?” hedads and mums simultaneously with bill, most open it; some have a query and call continues. “my kids weren’t taught thattargeted campaigns that fuelled interest. the centre. forty thousand bills go out an huge volumes of data were great. was as Laurence explains, it’s all about making hour but if the centre gets hit with too many there a university professor who stood upthe data work. “i triangulate an objective queries, billings are dialled down to reduce and said, ‘if you have over 100 indicatorsassessment of the new technologies coming calls in. it’s about fighting the data overload. you’re a good boy’? because whoever thatin, a subjective assessment of the public’s and we are truly overloaded by data. professor is, we need to shoot him.”reaction to new propositions, and then i take a Governments around the world are Laurence has just won a wager with hispunt.” this ‘triangulation’ is the combination unleashing a tsunami of numbers on their team over the number of Vodafone Vipof hardheaded data analysis, coupled with citizens. that has huge implications for members who bought tickets for concerts.business nous. data is something that big businesses with lucrative government his team, based on the data, bet on oneinforms his hunches – but never rules them. contracts. in the uK, the government number. their boss, based on what he setting up the £5 million network recently published every item of public knows about people, thought it would beOperations centre (nOc) in newbury spending over £25,000. search the database higher. data plus hunch equals a powerfulwas the first expression of this approach at for ‘Vodafone’ and you get 2,448 individual combination. Or, as Laurence concludes:Vodafone. “it’s very difficult to touch and transactions covering millions of pounds. “data on its own is impotent.” 12
What is your earliest memory? What do you Want Who is your inspiration? that you can’t have?pouring corn flakes into a bowl at the anita roddick from the body shop. rip.age of four without asking permission a teleporter.from my mum. What Was your greatest mistake? When did you last feel ashamed?What’s your signature dish? not taking enough risks. i tried to do my 15-year-old daughter’sGiven my cooking abilities, signing the chemistry homework and failed.bill in a restaurant. What gets you out of bed in the morning? What are you searching for?if you had to stay in one an alarm clock and the fear of being inplace, Where Would it be? my ipod. One of the kids has borrowed it. the house when the kids wake up if they went to bed late the night before.the colombe d’Or restaurant in saint-paul de Vence, france. When Were you last surprised? Which piece of music Last week. a customer sent me some alters your state of mind?When Was your last cookies to say thank you for showing themmoment of clarity? around our network Operations centre. in a positive sense – Lady Gaga at 7am if everything goes wrong in my current on the m4. in a negative sense – wagnerthe last time i spent time with a career i might become a tour guide. at any time on any motorway.customer. fortunately, that’s quite often. What is your greatest What do you Want toWhat does success extravagance? be When you’re older?look like to you? Quality wine. i don’t play golf, go to the about 10kg lighter.spending an hour on it, and not being pub, stay out late with the lads or go toable to improve it. casinos, so the deal with my wife is that she doesn’t ask how much the wine costs. tell us a joke...What is your biggest failure? i like simplicity in life. i heard this What do you see in the mirror? urban myth a long time ago and iti agreed with myself that i would get stayed with me. when nasa firstfit as soon as things calmed down at pierce brosnan on a good day, started sending astronauts into space,work. we had the conversation in 1982 Jeremy clarkson on a normal day they quickly discovered that ballpointand i’m no further forward. and Quasimodo on a bad day. pens wouldn’t work in zero gravity. to combat the problem, nasa scientists spent a decade and $12 billion developingWhen did you last let yourself go? hoW much is enough? a pen that writes in zero gravity, upside down, underwater, on any surface and atcreatively, whenever i can; financially, i don’t know yet, but i promise to tell temperatures ranging from below freezingnever; mischievously, every day. you when/if it happens. to 300°c. the russians used a pencil 15
A DATA STATE oF MIND interview by Ulrike reinhard portrait by erika SVenSSOni n fo g r a p H i c s b y M O r i t z S t e fa n e r data superstar H a n s r os l i n g explains wHy a fact- bas e d wo r l dv i e w w i l l t r a n s fo r m yo u r b us i n es s . 16
80 70 60 50 40 30 Average life expectancy after birthLIFE EXPECTANCY 20 10 01 02 03 04 05 06 07 FERTILITY RATE Fig. 1 Average number of children per woman 18
How did a public health official from understand what is happening in terms ofsweden become the world’s most famous potential new markets in the middle east,statistician, a television personality and a africa and so on, they are out. so the biggerregular guest speaker at corporate events? and more international the organisation, as an undergraduate, Hans rosling the more fact-based the ceo’s worldviewstudied statistics and medicine at uppsala is likely to be. the problem is that they areuniversity, sweden. He earned a phd, spent slow in getting their organisation to follow.two decades studying in africa and, aschairman of the Karolinska international tq why is this?research and training committee, hascollaborated with universities in asia, rOSling companies as a whole areafrica, the middle east and latin america. stuck in the rut of an old mindset. they throughout his career, rosling has think in outworn categories and followmaintained a fact-based worldview – an habits and assumptions that are not, orunderstanding of how global health trends only rarely, based on fact. they need toact as a signifier for economic development break out of that to understand the worldbased on hard data. today, he argues, the way it really is. for instance, in termscountries and corporations alike need to of education levels, we no longer liveadopt that same data-driven understanding in a world that is divided into the westof the world if they are to make sense of and the rest; our world today stretchesthe changes we are experiencing in this from canada to yemen with all the othernew century, and the opportunities and countries somewhere in between. there’schallenges that lie ahead. alongside his son ola and daughter-in- “My basic idea is a broad spectrum of levels and we have to realise that asia, brazil, latin americalaw anna, rosling created the gapminder that the world and, to some extent, the middle east arefoundation to facilitate that process.using trendalyzer (a bespoke software has changed catching up with the countries we used to call the ‘west’.tool later sold to google), the roslings so Much, what but even when people act within ahave reinterpreted static health data as people need isn’t fact-based worldview, they are used tomoving, interactive graphics. the results talking with sterile figures. they areare revelatory, bringing a new awareness More data but used to standing on a podium, clickingof the social and economic history of a new Mindset.” through slide shows in powerpoint ratherglobal health, and demonstrating that than interacting with their presentation.creative applications of data can yield the problem is that companies haveextraordinary results. a strict separation between their it department, where datasets are produced,think qUarterly you’ve long been and the design department, so hardly anya proponent of hard data and statistics. presenters are proficient in both. yet thisin what sense do ceos need to change is what we need. getting people used totheir mindset in order to develop a more talking with animated data is, to my mind,fact-based view of the world? a literacy project.hanS rOSling my basic idea is that the tq what kind of data should we beworld has changed so much, what people looking at to gain this new mindset?need isn’t more data but a new mindset.they need a new storage system that can rOSling what’s important today is nothandle this new information. but what just financial data but child mortalityi have found over the years is that the rates, the number of children per women,ceos of the biggest companies are education levels, etc. in the world today, it’sactually those that already have the most not money that drags people into modernfact-based worldview, more so than in times, it’s people that drag money intomedia, academia or politics. those ceos modern times. i can demonstrate humanthat haven’t grasped the reality of the world resources successes in asia through healthhave already failed in business. if they don’t 1980 2009 being improved, family size decreasing 19
1800and then education levels increasing. rOSling we found that the mostthat makes sense: when more children important thing when presenting our datasurvive, parents accept that there is [on graphs such as the Health and wealthless need for multiple births, and they of nations, which tracks 200 years of globalcan afford to put their children through life expectancy versus income per personschool. so pfizer have moved their in a four-and-a-half-minute video] was notresearch and development of drugs to to put time on the x-axis. we made timeasia, where there are brilliant young move, and when you see the movement,people who are amazing at developing the data becomes like a football matchdrugs. it’s realising this kind of change – you can see who is catching up or, forthat’s important. instance, that a country like bangladesh that’s why ceos ask me to talk to is reducing its child mortality rate fastertheir staff – so they can learn to look at than sweden ever did [see fig. 2].these interactive videos and gain this bangladesh is still at a low levelnew mindset. then they’ll realise what economically, but at the same time therehas changed. in my first ted talk in 2006, is a huge internal market with cheapi made al gore get up on stage. i showed distribution and only one language.that in vietnam today they have the same so if you are a company with ambition,average family size as the us, and the you have to be in bangladesh. it’s one ofsame health as the us in 1980 [see fig. 1], the 10 biggest countries in the world, butand their economy is growing faster than people’s mindset leads them to believethe us. al gore told me, “i didn’t have that bangladesh is a hopeless placethe slightest idea.” the problem isn’t that in need of aid. what is so strong withspecialised companies lack the data they animation is that it provides that mindsetneed, it’s that they don’t go and look for it, shift in market segmentation. we canthey don’t understand how to handle it. see where there are highly developed countries with a good economy andtq How has gapminder managed to a healthy and well-educated staff.present data in such a way that you’re ableto change people’s preconceptions? tq are there any points of resistance 20
fig. 2 CHILd moRTALITY 0–5-year-olds dying per 1,000 born Bangladesh | Sweden 250 200 150 100 50 0 2010to this process of shifting people’s a clear division of labour between thosemindsets? who provide the datasets – like the world bank, the world Health organisationrOSling at the moment, i’m or companies themselves – those whoquarrelling with sweden’s minister of provide new technologies to access orforeign affairs. He says that the west process them, like google or microsoft,has to make sure its lead over the rest of the and those who ‘play’ with them and giveworld doesn’t erode. this is a completelywrong attitude. western europe and “the probleM data meaning. it’s like a great concert: you need a mozart or a chopin to writeother high-income countries have to isn’t that wonderful music, then you need theintegrate themselves into the world in thesame way big companies are doing. they specialised instruments and finally the musicians. meteorologists are one group that hashave to look at the advantages, resources coMpanies lack a ready grasp of this idea. they receiveand markets that exist in different places the data they a huge amount of data, which theyaround the world. process in a highly sophisticated way, and some organisations aren’t willing need, it’s that translating it into stunning graphics –to share their data, even though it would they don’t go and there they are on prime-time tvbe a win-win situation for everybody andwe would do much better in tackling the and look for presenting the weather while we all watch. this is exactly what we strive to emulate.problems we need to tackle. last april, it, they don’t we want our economic indicators, ourthe world bank caved in and finally understand how social indicators and our environmentalembraced an open data policy, but the indicators to be communicated onoecd uses tax money to compile data to handle it.” prime-time television with the same leveland then sells it in a monopolistic way. of efficiency.the chinese statistical bureau provides this is what we’re trying to do at thedata more easily than the oecd. the gapminder foundation – and this is whatrichest countries in the world don’t have ceos want their employees to do – playthe vision to change. with data and give it meaning i call this the ‘database huggingdisorder’. to heal it, we have to instil gAPmINdER.oRg 21
Words by Sarah BrackingI L L U s T r AT I o N s b y M at t tay l o rCAN yoU do bUsINess WhILe doINg good INThe deveLopINg WorLd? The ANsWer Is yes, bUToNLy If yoU foCUs oN The dATA ThAT mATTers.Investing in the developing world is opportunities like this are not onlyback at the top of the business agenda. manifold and potentially lucrative; whenAnd it’s about time, too. emerging carried out responsibly they can also actmarkets, including the burgeoning as a catalyst for transformation. sociallyopportunities presented by some African responsible investments won’t just leadstates, are fuelling an upsurge in interest to private gain – they have the potentialfrom the private sector. for example, to shape the world.in 2009, Angola registered a 109 per but the power to effect change is acent rise in foreign direct investment double-edged sword; poor investments,(fdI) as a percentage of gross fixed irresponsibly made, will have just as widecapital formation. As a whole in 2010, an impact – only this time it won’t be fordeveloping and transition economies the greater good. When the ripples ofattracted half of global fdI inflows, investment seep beyond private borders,‘leading the fdI recovery’ according questions of risk and the potential for lossto the United Nations Conference on or gain become a global affair. Lives, notTrade and development (UNCTAd). just bottom lines, are potentially at stake. 23
he historical challenge of investing inpoorer parts of the globe was magnifiedby widespread decolonisation in the1960s and ’70s. In Africa, expropriationof foreign-owned enterprises forcedbusinesses to rethink whether it wasnecessary to own rather than maintainsecure access to local assets. over thenext 40 years, companies lengthenedtheir supply and subcontracting chains,and generated new types of relationships,from parallel investing with publiclyfunded development finance Institutions(dfIs), to using state-backed exportcredits, and developing other risk-sharingrelationships such as leasing, forward-contracting and investment agreements. however, the instinct that having a stakein a derivative income stream from an assetin the developing world is safer than anownership stake in the actual fixed asset hadto be reassessed after the global financialcrash of 2007-8, since so much derivativestock was proved worthless or degraded.At the same time, such arms length contactwith developing countries has often provedof little use to the countries themselves,sometimes provoking a backlash againstoffshore equity, as happened in buenosAires, which banned all investments fromshell companies held in tax havens in 2005. 24
so how does one go about making sound amongst those riddled with risk? The most to be a good predictor of transparency,and socially responsible investments in common indices for investment risk are which in turn is positively related tothis new era? by focusing on the facts. the International Country risk guide from economic growth. The most recent World The poorest countries often present the New york-based prs group; ratings bank governance Indicators measurethe most challenges to today’s global from the economist Intelligence Unit, the quality of political institutions acrossinvestor, not least because future risk and from eurasia group; alongside the six categories: voice and accountability,is highly context-specific. Assessing more traditional standard & poor’s, dun political stability and absence of violence,diverse and fluctuating contexts is & bradstreet and payments data from the government effectiveness, regulatorygenerating ever-greater complexities of bank of International settlements. quality, rule of law and control of corruption.data, and bringing the worlds of business These indices focus on two aspects ofand academia – particularly political political risk: regime (in)stability and the (un)scientists and international development certainty of the macro policy environment. Development Dataexperts – closer together. digesting that but these indicators struggle to keep pace Making the numbers work for you.data, however, is another story entirely. with nebulous political identities and regime business data, economic, social characteristics. The predicted ‘top 10 most Almost every aspect of a country’sand governance indices, corporate dangerous countries’ at the beginning of a socioeconomic reality is documentedsocial responsibility measures and decade are rarely the same 10 that actually by statistics. but filtering through thedevelopment impact data are combining collapse by its end. evidence of political abundance of information and retrievinginto effective predictive instruments. stability on its own is not enough: while a reliable dataset that answers questionsbut hazards remain, not least in the level authoritarian regimes can prove safe places about the world we live in is not always asof mathematical complexity generated. for investment for a time, sometimes a long straightforward as it seems.Is the world really this complicated, or is one, a lack of democracy means that change, over 175 governments send data tothe data industry out of control, feigning when it does come, tends to be eruptive and the Imf for the International financialprecise forecasting but exhibiting no unpredictable. businesses can quite literally statistics and to the World bank for Worldgreater reliability than gut instinct? lose everything. development Indicators, which in turn The only way through this statistical so how can predictive datasets are used by investment risk analysts. Theblizzard is to look at the figures, one like these be improved? for starters, most well-known development indices,dataset at a time. institutional quality must be taken into the human development Index and account, since political risk ‘events’ (i.e. the data collected for the purposes of mass protest or regime change) are better checking progress towards the millenniumpolitical Risk understood by knowing how political development goals (mdgs) provideCan it be measured? institutions are likely to react. some can an overview of wealth and wellbeing in manage rapid change, while others falter. developing countries. It may not be obviousThere was a brief period in the 1950s when Institutional quality measures, which are at first how we can use this knowledge tothe boundaries of the Cold War defined proving reliable, will mark the future world, make better, more effective and ethicallya space that Western governments were and it is in the contemporary design of sound business choices, but even aprepared to protect for business. but that these that academics and business people perfunctory analysis shines some light. forworld is gone. In its place is a complex are meeting. for example, the freedom example, secondary school enrolments frompolitical geography where the apparent house index, Transparency International’s World development Indicators are a goodstability of a country can change quickly, Corruption perceptions Index (CpI), or sign of a more productive workforce, ofas has been the case in Tunisia or egypt. the quite specific polity Iv series all depict better governance and the probability in turnConversely, countries widely considered institutional quality, and provide context of a stable macroeconomic environment.dangerous, such as the eastern Congo, and depth to investment and risk planning. These early development indicatorsAngola, myanmar/burma, or sudan, The CpI is good for general context, but have recently been complemented by moreare proving profitable for business – as relies on perception, which can lag behind complex impact assessment tools – usuallyChinese and Indian corporations have actual political change. freedom house has in response to demands from the public ordiscovered, leaving the rest of the world only a handful of classifications: free, partly donors concerned about a project’s widertrailing behind. free and unfree. but the polity Iv gives impact. Concern with carbon emissions, or faced with this unpredictable landscape, accurate measures of the legal constraints assessment of an investment’s impact onhow does one spot a sound investment on a country’s executive, which turns out culture, heritage or happiness might seem 25
irrelevant to the businessperson of the past, und entwicklungsgesellschaft (deg) and can prove corporate social responsibilitybut future leaders will not be able to evade the emerging markets private equity and defend against the risk of reputationaldemands for quantifiable, evidence-based Association (empeA) framework are damage affecting the customer base.statements. In this way, data can help us both exemplary systems that measure the ‘Clean’ goods are in demand, and whetherachieve more transparent and accountable impact of a dfI investment. it’s diamonds vetted by the Kimberleyworking practices. Adopting measures like these aren’t just process or cocoa trading structures by but how do we value the quality of air, the about ‘being good’. According to advocates The fairtrade foundation, ethically soundprotection of a heritage site, the treatment of corporate social responsibility, meeting production practices can all be corroboratedof workers? An overall assessment of a the ‘triple bottom line’ of financial, social by datasets. Likewise, guaranteeing ancompany’s ‘social worth’ is some time away, and environmental returns – or ‘people, associational distance from child labour,but those taking the bridgehead approach planet, profit’ – grows the business in the environmental harm and land grabbingwill be thinking about this now. efforts long term. Customers are increasingly is just as imperative, though they do require ain ‘greenwash’ will no longer satisfy the demanding that their money be put to good much more sophisticated dataset.informed global public of the future. use; that the businesses they choose to buy All this information already exists. so how can intangible externalities that from promote democracy, social welfare and It can help you do business and ‘be good’.affect social welfare or the environment development on the ground – or, at the very Learning how to filter the noise and focusbe measured? An active relationship with least, don’t reverse patterns of progress. on the facts that matter to you is the firsta development finance Institution is a for example, widespread hIv awareness, step. Then an even bigger issue comes intogood place to start. such as that sponsored by Aureos Capital play: how do you make sense of it, digest it The dfIs are required to produce (with investment from CdC group and and absorb its meaning into the work thatmatrices of developmental impact, which Norfund) can reduce hIv prevalence, you do?means they demand more social value which can be measured by the World bank’s The answer is simple: you need tofrom their private sector co-investors. for development Indicators. successes like this visualise what the data is trying to say.example, the Corporate-policy project make customers happy. The four organisations below are doingrating (gpr) of the deutsche Investitions- better use of development indicators just thatGapminDeR ‘make data more actionable through design’. Watch to publish a constantly updated mapgapminder.org With a focus on international development, that charts relative poverty and wellbeing they help government agencies and the across the world. This easy-to-use tool hasIf you think global statistics are boring, think private sector embrace the open data helped everyone from the UN to Dell realiseagain. Gapminder’s bubbly Trendalyzer tool revolution by making complex datasets easy that evidence-based decision-making canbreathes life into the trends shaping our world. to understand. Quirky toolkits like MapBox be a pleasure, not a chore.By reimagining obscure patterns of social and Managing News turn tough data intochange as graphs that move organically easy-to-read visualisations or maps, andover time, Gapminder is smashing through have been used by everyone from Google to HealtHmapthe mythical glass ceiling that hangs over the World Bank. Whether they’re processing healthmap.orgthe ‘developing world’ and inspiring more election results in Afghanistan or monitoringpeople to absorb the facts. Everything from relief efforts in Haiti, simplicity is key. If the campaign for open data needs awealth and health to education and climate poster child, then HealthMap is it. Thisis rigorously analysed, then effortlessly online mapping tool aggregates informationinterpreted as dynamic graphs that represent statplanet from disparate open data sources to offer alife in every corner of the globe. sacmeq.org/statplanet comprehensive view of the state of global health. This year sees the launch of Predict, a Free to download, StatPlanet is a browser- tool that will help the public track outbreaksDevelopment seeD based application that creates customised of animal diseases that might affect humans.developmentseed.org maps, graphs and visualisations from all Pooling information from sources such as the manner of interlocking datasets. It’s turned World Health Organisation, Google NewsThe straight-talking brains at Development Transparency International’s befuddling and the Wildlife Disease Information Node,Seed have created an innovative range of Corruption Perceptions Index into an HealthMap proves that freely availabletools to combat information overload and interactive gateway, and allowed Social information can be a progressive social force. 26
THEKNOWLEDGE sIMoN rogErs PICks THE 10 BEsT PlACEs To sEE ‘sExY’ DATA oNlINE. Flowing Data flowingdata.com OWNI owni.fr If you’re looking for time series economic data – and a nifty way of creating a sophisticated, embeddable graphic – this is the place to come. Timetric updates thousands of datasets every day and provides an easy-to-use interface that makes it very simple to create your own. Timetric timetric.com If someone, somewhere, is producing Information a great data visualisation or analysis, Nathan Yau’s blog will find it. Yau has is Beautiful an unerring ability to unearth the best data visualisations on the web. He also produces graphics, and is a regular poster informationisbeautiful.net to the Guardian Datastore Flickr group. Canadian Patrick Cain is a ‘journalist who makes maps for the web’. Based in Toronto, Cain takes the city’s data and maps Patrick Cain’s it – producing guides to everything from crimeData journalist and design whizzDavid McCandless’ Information isBeautiful blog is a treasure-trove of cool Map Blog figures to World War I deaths and single parent families. A fan of open patrickcain.cavisualisations and mash-ups. His work data, Cain has a recordhas also been published in a bestselling of demanding data frombook of the same name, and we’ve got the city’s authoritiesa bunch of copies to give away. Email using Freedom firstname.lastname@example.org to get one. Information laws. 28
Although a lot of the best data work is This brand new sitedone in English, Paris-based oWNI is combines an innovativea collective of geeks and data freaks data search function with DataMarketproducing visualisations and apps that bright and imaginative visualisations. It alsomanage to be imaginative and innovative. allows you to create yourThe collective’s work on Wikileaks – own, download them datamarket.comwhich allowed people to interrogate the and put them in yourdata – won a 2010 online Journalism PowerPoint presentationAward for general Excellence. or company report. LinkedIn linkedin.com Guardian Datablog guardian.co.uk/data It might be better known for its impact on the world of social media, but linkedIn also has a hugely innovative approach to data. linkedIn has made collating and using data a priority, with lead data scientists completely integrated into the commercial operation.The Guardian and its Datablog publishes raw data behind the news Londonevery day, and encourages readers to visualise and work with it. Thesite publishes its data using google spreadsheets and google Fusion DatastoreTables, and allows readers to search thousands of government datasets data.london.gov.ukaround the world.The big brains atInfochimps have comeup with an innovativeway to find, share andsell formatted data. Both governments around the globe are opening up their Infochimpsusers and the site’s own data, from data.gov in the Us, via Australia, the Uk, Newcontributors collate and Zealand and France. one of the best and most useful isscrape datasets so that the london Datastore. Created by the greater londonthey’re easily accessible. infochimps.com/datasetsWith big plans for Authority, it publishes thousands of datasets with theexpansion and lots of emphasis on useful, live data, such as transport andintellegegent developers economic numbers. Developers are using those figuresonboard, it’s definitely to create interesting apps, such as Matthew somerville’sone to watch. live train map for the london Underground 29
Lunch With haL Hal Varian, CHief eConomist at GooGle, sinks His teetH into data obesity and How to treat it. ot 10 minutes into our lunch at Google HQ in Mountain View, California, a groupie sidles up. He’s got a guest nametag and an outstretched hand. He wants to say hello to Hal Varian, Google’s Chief Economist. Varian, employee number 441, author of Information Rules, Emeritus Professor at the University of California, Berkeley (where he was founding dean of the School of Information), and a former columnist for The New York Times is, it turns out, quite a star in the statistics world. The young man quotes a line from the elder circa 2009. That’s when Varian famously pronounced that “the sexy job in the next 10 years will be statisticians.” He added: “I’m not kidding.” Of course, sexy, like funny, is subjective (does a T-shirt that reads ‘Statisticians do it with models’ make you laugh?). But Varian’s prediction is backed up by trends. In 2010, the human race created 800 exabytes of information, from tweets and Facebook updates to PowerPoint presentations and photographs. That’s 800 billion gigabytes, or the amount of data you can fit on 75 billion 16-gig iPads. To put that into context, between the dawn of civilisation and 2003, we only created five exabytes; now we’re creating that amount every two days. By 2020, that figure is predicted to sit at 53 zettabytes (53 trillion gigabytes) – an increase of 50 times. Multiply data and you multiply the need for people to make sense of it. That’s where Varian and the statisticians, analysts and econometricians who work with him come in. wo r ds by H o l ly F i n n i l l U s t r at i o n s b y a d r i a n j o H n s o n31
Data is like food, says Varian. “We used year, while looking to buy a pepper shaker real-time database. And that’s powerful.”to be calorie poor and now the problem online, he hit upon the idea of a Google Government can, in turn, aggregateis obesity. We used to be data poor, now Price Index (GPI). It uses Google’s web information, giving businesses insight intothe problem is data obesity.” Google’s shopping database to create a daily their industry and the economy as a whole.strength, he continues, was to recognise measure of inflation. It could, one day, The trick, in both directions, isback in 2001 that “we would be handling be a complement – or competitor – to the getting high-quality data. But neithermassive amounts of data, and would official, yet less frequent, Consumer Price governments nor businesses guaranteeneed to develop tools for that.” Another Index (CPI). it. Last year, Canada’s conservativeforesight was to hire an economist. There’s a systemic gap, Varian points cabinet voted for a weakened census.Eric Schmidt hired Varian to ‘have a look out, between the low-frequency data By removing the requirement for citizensat the auction’, the bidding system for employed by governments and the high- to fill out a ‘long form’ (consideredads that soon became Google’s lifeblood. intrusive), it reduces its own access toOther companies were built up around vast amounts of quality information.auctions – eBay, Yahoo! – but they didn’t “The head of the census bureau resignedhire experts until much later. Now over it,” says Varian. “And there’s someYahoo!, Microsoft, Apple and Intel all discussion of similar things here [that]have chief economists. the US is considering.” For businesses that are gorging on a Companies, too, should be concernedsurfeit of information, Varian says the fix “if yoU’re tHe about the quality of the decision-is clear. It’s the same for data as food: “Youneed to focus on quality. You’ll be better CHief anytHinG, influencing data they are getting, says Varian. He recalls Lou Gerstner, beforeoff with a small but carefully structured yoU always HaVe joining IBM, doing reconnaissance.sample rather than a large sloppy sample,” a problem witH Externally, “he asked people how thehe says. More locally sourced fine dining, company was doing, and everyone gave itthen, less all-you-can-eat buffet. people tellinG a C. Then, when he got to the company, Varian looks trim enough, dressed yoU wHat yoU he asked the same question: ‘How’rein a blue shirt and plain khaki trousers,with brown shoes and a navy sleeveless want to Hear. we doing?’ And the answer was: ‘All our customers give us an A!’ So he said,sweater – the uniform of a mind with more it’s Hard to Get ‘Where’s that data from?’ ‘Oh, we askedimportant things to think about than CritiCism.” our sales people to collect it.’fashion. He takes a similarly practical “If you’re the chief anything, youapproach to his food. always have a problem with people This on-campus café serves a telling you what you want to hear,”smorgasbord of exotic treats – including, adds Varian. “It’s hard to get criticism.”today, Beautifully Braised Short Ribs, So Varian courts it. His team conductsLocal Oysters (Kumamoto, Point Reyes random surveys constantly. Theand Marin Bay), Artichoke Poached in Ad Happiness Survey, for instance,Court Bouillon, and Espresso Chiffon frequency data of business. Government continually picks advertisers at randomCake. But Varian arrives at the long is working on it, though. “It’s now using and asks them about their experience. Inwhite refectory table having quickly supermarket scanner data to predict order to remove some selection bias, he’sheaped his plate with iceberg and inflation rates,” notes Varian. How did it even hired someone to be (at least someshredded carrot doused in Thousand predict them before? “They used to send of the time) ‘Chief Nag’ to get results outIsland dressing. He’s also got a hastily people out with notebooks to write it down.” of resistant responders.plonked-together chicken salad More communication between Google also conducted 5,000 searchsandwich on a white roll with one slice government and business clearly benefits experiments last year, which led toof tomato. You can take the boy out of both, says Varian. Business can provide 400 search improvements (and theWooster, Ohio… more real-time data. “If you look at most same again for ads experiments). Such For Varian, everything – including his businesses now, pretty much everyone – insistent experimentation is an academicculinary choices – can relate to data. Last think of UPS, FedEx, MasterCard – has a means to a capitalist end. Or, as Varian 32
puts it: “Google is like a university, butwith money.” It’s a cheering thought – andpossibly why, in a recent talk at the 150thanniversary of MIT, Varian was notablymore optimistic than his peers. “Economics is really on a roll inSilicon Valley,” he says. “The goodnews is that standard techniquesfrom economics work very well onbig problems. It’s a little discouragingworking as an academic economistbecause the problems that you work onare so hard. They’re much, much easierin industry.” He means all industry, not justFortune 500 companies. Varian, aman who quotes feminist playwrightEdna St. Vincent Millay as easily asMIT founder William Barton Rogers,points to the proliferation of what hecalls ‘micro-multinationals’ – smallcompanies, mostly connected touniversities, working around the globeand around the clock. “The smallestcompany now has access to computingand infrastructure only the biggest techcompany had 15 years ago. So it’s a muchmore fertile environment for start-ups.We’re seeing all these little companies.And guess what, some of these littlecompanies become big companies.” Information Rules, Varian’s seminalbook with Carl Shapiro, re-popularisedthe phrase ‘network effects’ – the value ofa product or service increasing as morepeople use it. His next book will focuson new ideas, including ‘co-opetition’,the notion of capitalist symbiosis(Google and news organisations bothchampioning content, for instance).Varian is always thinking about what’snext. Asked to define his job, he says: “Toanswer the questions that managementwill ask next month.” So, what is next for the economy?Is it gaining steam? “Yes,” says Varianemphatically. And how does he know?“I inspect the entrails,” he says. Foodis data, and data a kind of food 33
A vISUALhISTORy OF DATACApTUReThROUghThe AgeS Why not open me up and check me out?
Tally sTicks 20,000-10,000 BC The Ishango bone, a tally stick from the Upper Palaeolithic era, represents the beginnings of our understanding of mathematics and data. sundials 3,500 BC Egyptian obelisks show humans manipulating light and shadows to measure data about the time of day. The book on PaPyrus numbers and 3,000 BC comPuTaTion 200 BC Papyrus, manufactured in Dating back to the Han Dynasty Egypt, revolutionises the of ancient China, this mathematical way data and language treatise brings together interest rate can be recorded. calculations with government statutes and law reports. census abacus 800-500 BC 2,700 BC In Israel, a primitive Ancient civilisations develop census is undertakena counting system that enables and recorded in the complex data manipulation. Hebrew Bible. Social data capture is born.
The domesday sTock book exchanges 1086 AD 1200s AD William the Conqueror The earliest stock exchanges conducts a survey in emerge in Bruges and Italy in England and Wales recording the thirteenth century. Data about land and livestock. It takes trades is written down by scribes over a year to complete. and transported by couriers. gregorian calendar 1582 AD navigaTional comPass Pope Gregory XIII launches the Gregorian calendar to 1000s AD eradicate an 11-minute discrepancy Chinese scientists develop in the Julian calendar, which is instruments that attract causing the official date of equinox a needle north, creating to creep further away from the a navigational tool only actual cosmological event. recently superseded by GPS. ThermomeTer 1600s AD Cornelius Drebbel, Robert Fludd, Galileo Galilei and Santorio Santorio make progress on a device to measure temperature in real time. TelescoPe analyTical 1600s AD engine 1837 AD Scientists in the Netherlands develop a refracting telescope Charles Babbage developsthat Galileo improves in subsequent the Analytical Engine, years. The instrument observes and modern computation remote objects in real time. is born.
wireless Telegraph 1897 AD Guglielmo Marconi founds The Wireless Telegraph & Signal Company, pioneering communication between coastal radio stations and ships at sea. DaTa VisualisaTion Telemobiloscope 1857 AD 1904 AD During the Crimean War, Christian Hülsmeyer uses Florence Nightingale records radio waves to detect distant the mortality rates of British metallic objects, inventing the soldiers in field hospitals. The first radar application. information is published in a series of striking graphics, persuading the government to improve conditions. Telegraph gps 1837 AD 1957 AD The first commercial Sputnik – the first artificial satellite –telegraph is introduced at Euston is launched by the Soviet UnionStation. It soon crosses the oceans on October 4, 1957, as a globalto every continent but Antarctica, positioning system for precise making instant global weapon delivery and paves the way communication possible for GPS as we know it today. for the first time.
google earTh engine 2010 AD The Google Earth Engine – a cloud computing platform – processes real-time satellite imagery and other Earth observation data. Initial applications of the platform include mapping the forests of Mexico, identifying water in the Congo basin, and detecting deforestation in the Amazon. clusTer personal exploraTory 2008 AD compuTer 1970s AD Cluster Exploratory (CluE) is a National Science Hewlett-Packard introduces Foundation-funded programprogrammable computers that that analyses massive amounts fit on top of a desk. The of data to search for patterns. personal computer allows economical collection and management of data. supermarkeT meTrics 1995 AD Tesco’s Clubcard scheme revolutionises consumer metrics by allowing supermarkets to target offers and optimise their stocks.raDio-frequency hubble space iDenTificaTion Telescope 1980s AD 1993 ADRadio-frequency identification The Hubble Space Telescope technology (RFID) takes hold captures images of outer spacein transportation and business. in real time, allowing scientists Real-time monitoring systems to determine the rate of are developed to process the expansion of the universe. new data.
Online videO advertising is allOwing brands tO speak tO audiences On a glObal scale. big ideas will reap rewards, prOvided yOu get tO knOwyOur audience by putting data first. wOrds by Ulrike reinhard very minute, 35 hours of footage is uploaded to youtube globally. with over two billion views a day, it’s become the epicentre of a video advertising boom. last year saw brands embracing innovative online video campaigns like never before. tipp-ex’s ‘shoot the bear’ and french connection’s ‘youtique’ led the way, generating millions of views and acres of publicity. these success stories are telling marketers that big numbers are within reach. but how do you go about creating a campaign capable of capturing a mass audience? is it through deep data analysis, or could the secret be something less tangible? as the following case studies with tipp-ex andfrench connection show, real-time data analysisduring and after the process, combined with anunexpected and interactive narrative, are thefoundations on which a campaign can be built.
hoot the bear was tipp-ex’s web clicking on the video and increase viewing debut. the briefing given to numbers,” he continues, “we first had buzzman, the paris-based creative to analyse code and implement certain agency behind the campaign, technical solutions within youtube’sTIPP-EX defined its goals as: ‘to raise short- term brand awareness and to be on top of customers’ shopping lists. to guidelines. this is where the ‘nsfw’ in the video title comes in – it stands for ‘not safe for work’. we checked all the most viewedShoot the Bear go europe-wide and tell the story of how videos on youtube. we analysed people’s the product is used.’ surprisingly, going behaviour, and when we examined all this digital wasn’t part of it, but after mining data we were pretty sure that we didn’t data on youtube’s most popular videos, want a branded video or our own tipp-Client BIC buzzman came up with a viral ad titled ex video channel. it became pretty clearAgency Buzzman NSFW. A hunter shoots a bear! that a simple video on the main youtubeSearch ‘NSFW. A hunter shoots a bear!’ in the 30-second clip, a hunter in a platform was the right thing to do. a video forest is approached by a bear. users are with the look and feel of a video shot on asked whether the hunter should shoot a mobile phone by you or me. the data the bear, and their decision leads to a taught us that we need to surprise the second video which sees the hunter reach viewer – and that’s what we did at the end out of the player to grab a tipp-ex pocket when the hunter starts freaking out. this Mouse from what appears to be a static ad is where interactivity kicks in. people love and erase the word ‘shoots’ from the title. being involved – being part of the story.” viewers are then invited to write whatever with a total layout of around €900,000 they want in the blanked-out space and (including production, advertising watch as the hunter does exactly what and agency fees), shoot the bear was a they’ve written. europe-wide campaign unbeatable in cost “we produced 42 scenes,” reveals efficiency. it went viral from day one: with thomas granger, Managing director at one tweet per second in the first 10 hours, buzzman, “with one search query for each and one million views after 36 hours. to scene. based on a survey, we found that date, the video has had almost 500,000 for each query – let’s say ‘plays with’ as an shares on facebook, been posted on more example – there were 40-60 words used by than 1,300 blogs and more than 43 million respondents to express the notion of ‘play’. people have watched it. so whenever somebody types in one of it’s a winner in business terms, too. a these expressions, the query leads them survey by tipp-ex showed that the ‘buying straight to the specific scene. real-time attention’ of potential customers – which data showed us which scenes were hot and positions the brand as the first product which were not. that’s a great source for they are likely to buy – increased by 100 identifying what youtube viewers want per cent, while sale volumes were up by 30 and telling us how to react. per cent compared to the same timeframe “to maximise the chance of people the year before. 36
nlike tipp-ex, fcuk is an old hand engage when the video set-up reflects at e-commerce. their goal was to their own lifestyle, rather than that of the grow business by reaching out catwalk jet-set. beyond their website and using “so our decisions weren’t by anyFrEnch connEcTIon new communication channels in an innovative way. not only would they reach customers in the us means based on blind judgments. data mining gave us clear indicators, which really helped in creating the rightYouTique and uk, but they’d also develop insights atmosphere in the videos. knowing the about the way video is used on the web. data, we were fairly sure that people poke, fcuk’s east london agency, would buy.” created youtique – a youtube boutique “since we made the experienceClient French Connection – as a place where youtube and ourselves, we were able to take measure-Agency Poke commerce intersect. it makes clever use ments and now we can optimise forSearch ‘French Connection, YouTique’ of youtube’s pop-up buttons by letting the upcoming season,” adds Jennifer viewers buy what they see with just a roebuck, director of e-commerce at few clicks. though the pop-ups tradition- fcuk. “the data tells us exactly the ally only link to other youtube videos, right length for our new videos, the fcuk was the first brand in the uk to best spots for calls for action, the best make an arrangement with youtube starting points, the best way to place to let them use what youtube calls content and label it to achieve number ‘annotations’, which enable viewers to one search results. these are just a few of leave the platform and go off to other the lessons the datasets are teaching us. destinations on the web – in this case to now we’re ready to improve.” the fcuk website. like the tipp-ex campaign, youtique “youtube data showed us that people is a winner in business terms, too. it was were actively searching and browsing among the three most popular youtube for fashion tips and tutorials, diy channels in the uk for a month. a 100 instructions that showed them how to per cent increase in channel views dress sexily for a date, what to wear on a meant a significant increase in brand business trip or what’s the latest fashion awareness while the click-through rates must-have,” says emma pueyo, creative were among the highest youtube has director at poke. “and the data also ever seen – up to five or six per cent – and showed that people are most likely to online sales soared. data-based web advertising was the big breakthrough of 2010, utilising the power of social media to transform the relationship between consumers and advertisers. tipp-ex and french connection have pioneered a new model. the next move is yours 37
AD VALUE Tony Fagan, DirecTor oF research aT google, answers The six quanT quesTions every cMo shoulD be asking in orDer To MaxiMise Their reTurn on search aDverTising. w o r D s b y T o n y Fa g a n i l l u s T r aT i o n s b y a D a M H ay E SWelcome to the age of experiments. whether something caused somethingAt Google, we believe that online else. In marketing, we call them ‘A/Badvertising is a more measurable medium tests’. The idea is simple: test A versus Bthan television, radio and print. How can to see which one works better. That giveswe be sure? Because we look at the stats. us the ‘incremental’ improvement – the Business is about trial and error, but difference between doing something andwith statistics comes a method to make not doing it.the process work better. With the data We can use these experiments togenerated from search, click-through and address six commonly asked questionsconversion rates, we’re able to address and about running search ad campaigns onimprove ad campaigns on the fly. Google. The answers will give you an The process is called ‘test-and-learn’ insight into how to make online advertisingand it’s the gold standard for calculating work efficiently for your organisation. 40