broadstuff

Ten Blindingly Obvious Things about PRISMGate et al

nospam@example.com (Alan Patrick) — Fri, 14 Jun 2013 21:46:45 +0100

Like many others I've been watching the whole PRISM issue unfurl with an increasing measure of amusement and amazement, mainly that people are surprised and shocked. There is so much BS being spouted in every direction, I thought it may help to remind everyone of the 10 Rules Of Social Data Mining:

1. "Data wants to be Free", and most people give it away as if it was. It may want to be free, but it is very valuable.

2. Everyone who can collect it therefore, will collect it.

3. Those that collect, will share or sell - the more data you have, the more valuable it is - there is a Metcalfe's Law of data.

4. That means Governments too, not just Google. Why were you surprised? Did you note that was Government s. Not just yours. Especially if all your data is on servers in another country.

5. They won't just collect it, they will store it. All of them. For a long time. A very long time.

6. They will all tell you it is all anonymised, and can't be used to find YOU. They are lying.

7. They will tell you they will only use it for specific things, not generic datamining, no need to worry. Worry. Remind yourself of (1) above.

8. They will tell you that if you have nothing to hide, you have nothing to fear. Fear. Yesterday's amusing peccadillo is tomorrow's thought crime.

9. They will finally tell you it is "necessary" to take it all, and store it forever - for cheaper services, for better efficiency, to win the War On Terror*. It isn't.

10. I will tell you I can see into your soul for the price of letting you put a few photos up and telling other people what you had for lunch. You better believe it.

If, while reading all the hoo-ha, you keep all that in mind, you may not lose your head

And (again) be careful what you put online. Datamining algorithms can tell a hell of a lot even from 30 days worth of your "what I had for lunch" tweet data once it's cross correlated with all the others in your network and teh other data out there that can be cross referenced to you

*They who can give up essential liberty to obtain a little temporary safety will get neither liberty nor safety - Benjamin Franklin

Iain Banks RIP

nospam@example.com (Alan Patrick) — Sun, 09 Jun 2013 22:03:46 +0100

Science Fiction writer (Culture series) and "Normal" fiction novelist Iain Banks (Wasp Factory etc) has died, aged 59.

Consider Phlebas, who was once handsome and tall as you. May he rest in peace.

Big Data vs Big Maths

nospam@example.com (Alan Patrick) — Sat, 01 Jun 2013 22:08:30 +0100

I'm liking Microsoft's Kate Crawford, she too is a Big Data sceptic - her "6 Myths of Big Data" in the NYT is exactly the sort of thing we would write, so we've copied it (expurgated) here, with a few comments [in brackets]. In essence she thinks that Big Data boosters (aka Fundamentalists) are labouring under the misapprehension that more data = more facts = more accuracy, and she has pointed out 6 myths around this:

Myth 1: Big Data is New

In 1997, there was a paper that discussed the difficulty of visualizing Big Data, and in 1999, a paper that discussed the problems of gaining insight from the numbers in Big Data. That indicates that two prominent issues today in Big Data, display and insight, had been around for awhile. “But now it’s reaching us in new ways,” because of the scale and prevalence of Big Data, Ms. Crawford said. That also means it is a widespread social phenomenon, like mobile phones were in the 1990s, that “generates a lot of comment, and then disappears into the background, as something that’s just part of life.” [Never mind 1997, Big Datasets have been datamined by Telcos - the first large social network owners - large retailers, airlines and bottom feedin credit card companies for several decades]

Myth 2: Big Data Is Objective

Over 20 million Twitter messages about Hurricane Sandy were posted last year. That may seem sufficient for a picture of whom the storm affected. However, the 16 percent of Americans on Twitter tend to be younger, more urban and more affluent than the norm. “Very few tweets came out of Breezy Point, or the Rockaways,” Ms. Crawford said. “These were very privileged urban stories.” And some people, privileged or otherwise, put information like their home addresses on Twitter in an effort to seek aid. That sensitive information is still out there, even though the threat is gone. That means that most data sets, particularly where people are concerned, need references to the context in which they were created. [And of course, Twitter is not the only source of large reams of data, the reason so many people focus on it is, like the drunk looking for his key under the street-light, it is far easier to look there]

Myth 3: Big Data Doesn’t Discriminate

“Big Data is neither color blind nor gender blind,” Ms. Crawford said. “We can see how it is used in marketing to segment people.” Facebook timelines, stripped of data like names, can still be used to determine a person’s ethnicity with 95 percent accuracy, she said. Information like sexual orientation among males is also relatively easy to identify. (Women are tougher to pinpoint.) That information can be used to determine what kind of advertisements, for example, that people receive.It’s important to remember that whenever people start creating data sets, these become fallible human tools. “Data is something we create, but it’s also something we imagine,” Ms. Crawford said. [And they are prey to all the same biasses that small dataset work is prone to - being big doesn't make it better - all data models are wrong, by definition]

Myth 4: Big Data Makes Cities Smart

“It’s only as good as the people using it,” Ms. Crawford said. Many of the sensors that track people as they manage their urban lives come from high-end smartphones, or cars with the latest GPS systems. “Devices are becoming the proxies for public needs,” she said, “but there won’t be a moment where everyone has access to the same technology.” In addition, moving cities toward digital initiatives like predictive policing, or creating systems where people are seen, whether they like it or not, can promote lots of tension between individuals and their governments. Sorry, IBM. Take that, Cisco. That goes for you, too, Microsoft, Ms. Crawford’s employer. All these big technology companies have Smart Cities initiatives. [Quite - as we have pointed out ad nauseam on this blog, freeing your data can be just another way of putting you in chains. But then, if you had nothing to hide, why wouldn't you want us to have your data, I hear you say.....]

Myth 5: Big Data Is Anonymous

A study published in Nature last March looked at 1.5 million phone records that had personally identifying information removed. It found that just four data points of when and where a call was made could identify 95 percent of individuals. “With just two, you can identify 50 percent of them,” Ms. Crawford said. “With a fingerprint, you need 12 data points to identify somebody.” Likewise, smart grids can spot when your friends come over. Search engine queries can yield health data that would be protected if it came up in a doctor’s office. [This is well known by anyone working with databases, but the Big Data fans - and open Data apostles - are strangely quiet about this issue]

Myth 6: You Can Opt Out

Last December, Instagram, the photo-sharing site, changed its terms of service to allow it to share customer’s photos more broadly, even use images in ads. What it didn’t have was a paid option, in which a person could, for a fee, not be part of that. Even if that option existed, Ms. Crawford said, this would imply a two-tier system — people who could afford to control their data and those who could not. [As we've argued before, the online world shows many worrying signs of devolving into a "feudal 2.0" system, with the digital serfs trading their data for access to crappy services relying on Ad funding, and you will only get truth by paying for it

In other words, Big Data behaves in much the same way as Not-So-Big Data really.

We'd also add a Coda - Statistics, and to an extent Operations Research (or Decision Maths or whatever the latest in-word is), is the science (or art, too often) of estimating what large data sets will contain from much smaller datasets, and once those sample datasets are above a certain size they are fairly indistinguishable from the overall dataset, so long as they are properly randomly sampled. A lot of the "insights" from Big Data - the "80/20" in my experience - are usually quite easy to glean from small datasets and "Big Maths". In fact, if I may be so bold, I do think a lot of the Big Data hoo-ha is from people whose main grasp of maths is spreadsheets with $ sign denominations.

We will definitely keep a closer eye on Ms Crawford's work. I suspect Microsoft may as well

McKinsey Disruptive Dozen Technologies

nospam@example.com (Alan Patrick) — Wed, 29 May 2013 11:21:59 +0100

McKinsey Disruptive Dozen Technologies

Hot on the heels of our look at McKinsey's Top 10 Technologies, we look at their "Disruptive Dozen" (They really have taken the "Number Of X" blogpost trope to heart ).

The first thing that hits me, looking at the chart above, is that some of these are relatively tiny in impact, so its unclear how they will be "disruptive" in any significant way. The other thing that hits me is that three of these tiny ones are new energy sources. This implies a huge discrepancy between the new energy source hype/expectations, and the likely reality. I am impressed that they have not been carried away with the hype around 3D printing, we believe it won't be that high impact either (see here)

Looking at the high impact technologies, the Cloud and Mobile internet are both already large, and their change vectors are well known. Existing systems have existed for quite some time, so why will these be disruptive? If you look at previous true disruptions, it usually comes from the "first off" delta with high penetration which has arguably already happened, not the later "build ons", in this case to the existing, globally available Server Farm/Web Hosting and Mobile Internet systems.

And as for "Knowledge Automation", this is supposed to remove the jobs of the 200 million Western knowledge workers. These, however are the people who are most connected and have political power. Besides, we suspect - again if the past predicts the future - that offshoring to lower wage economies is far more likely. Whichever, selling Automation and Globalisation was one thing when it was for blue collar workers in the boomtime, it'll be another thing when it hits lawyers, doctors and bankers in the Great Recession. History implies this will be a bunfight....

The areas that they do flag that we agree are both large and disruptive are the Internet of Things, and Advanced Robotics (of which autonomous vehicles are really just a subset). More on our thinking about these here and here.

They seem not to have featured one thing they did have in their Top 10, ie the next 3 biliion people in poor countries joining the internet via mobile systems - now history suggests that will be extremely disruptive across all vectors.

It's a bit Curate's Eggy in my view, no doubt to spur debate. Anyway, the report is well worth a read, and there is a livechat later today on the topic on #McKDisrupt on Twitter

McKinsey's latest 10 Tech trends - the Broadstuff view.

nospam@example.com (Alan Patrick) — Sun, 26 May 2013 21:54:00 +0100

Modified Gartner Hype Curve applied to McKinsey Forecasts

McKinsey's latest set of ICT trends. This is a follow on from their 2010 forecast (we reviewed that here). Here is the Broadstuff expurgated version, and probably one of the few where you are likely to get a bit of a qualified reality check [In Brackets]

1. Joining the social matrix

Social technologies are much more than a consumer phenomenon: they connect many organizations internally and increasingly reach outside their borders. The social matrix also extends beyond the cocreation of products and the organizational networks we examined in our 2010 article. Now it has become the environment in which more and more business is conducted. Many organizations rely on distributed problem solving, tapping the brain power of customers and experts from within and outside the company for breakthrough thinking.

[Its a transaction cost game - see our thoughts on this over here - and in very early days. It is also largely Just Another Channel, not a New Paradigm - that was so 2012]

2. Competing with ‘big data’ and advanced analytics

Three years ago, we described new opportunities to experiment with and segment consumer markets using big data. As with the social matrix, we now see data and analytics as part of a new foundation for competitiveness. Global data volumes—surging from social Web sites, sensors, smartphones, and more—are doubling faster than every two years. The power of analytics is rising while costs are falling. Data visualization, wireless communications, and cloud infrastructure are extending the power and reach of information.

[The problem with Big Data (or data analysis, as it used to be known in the Olde Days, c 2010) is that the low hanging fruit is soon plucked (if it hasn't been already - people used quaint old ideas like Statistics and Operations Research to estimate these larger datasets and optimise things in them thar Old Days). The other issue is Beautiful Mind Syndrome* where people see patterns in the data where none exist (or worse, you anchor on non patterns you want to see) and chase after will o' the wisps. Big Data is a lot like SEO - typically once you've done the fairly obvious things (that you may well have already done - see Statistics, use of, above) you can shave small incremental benefits off things by analysing them to the nines, but a crap business will still be a crap business and you're better off changing that than crunching yet more data]

3. Deploying the Internet of All Things

Tiny sensors and actuators, proliferating at astounding rates, are expected to explode in number over the next decade, potentially linking over 50 billion physical entities as costs plummet and networks become more pervasive. What we described as nascent three years ago is fast becoming ubiquitous, which gives managers unimagined possibilities to fine-tune processes and manage operations.

[We have been active in this area for 20 odd years, and the hype still outpaces the reality unfortunately. Unit costs per sensor and the cost of integration, lack of standards and reliability of these systems at any scale are still huge barriers to widespread deployment. But it is coming, one day...see our take on it over here]

4. Offering anything as a service

The buying and selling of services derived from physical products is a business-model shift that’s gaining steam. An attraction for buyers is the opportunity to replace big blocks of capital investment with more flexible and granular operating expenditures. A prominent example of this shift is the embrace of cloud-based IT services. Cosmetics maker Revlon, for example, now operates more than 500 of its IT applications in a private cloud managed by an external provider. It saved $70 million over two years, and when one data center in Venezuela was hit by a fire, the company was able to shift operations to New Jersey in two hours. Moves like this, which suggest that cloud-delivered IT can be reliable and resilient, create new possibilities for the provision of mission-critical IT through external assets and suppliers.

[The best X As A Service operators are making thin margins at best, and increasingly having to grow by acquisition (Salesforce.com), and most are still heavily subsidised by their parent companies or running on investors money. The hassle factor of running key services remotely, with the same level of reliability and flexibility as in house, is still not a slam dunk for anything but the very basic commodity services. Exits (and valuations) are a function of hype over reality still]

5. Automating knowledge work

Physical labor and transactional tasks have been widely automated over the last three decades. Now advances in data analytics, low-cost computer power, machine learning, and interfaces that “understand” humans are moving the automation frontier rapidly toward the world’s more than 200 million knowledge workers.

[This we will watch with interest, as those 200 million are the bulk of the educated elite in developed countries, they are connected and (still) wealthy, and are not going to take this development lying down. Business leaders were able to push Globalisation when it was boomtime and only low end blue collar jobs were being hit, lets see if it still works with high end white collar ones, in Great Recession hit democratic countries. Also, the automation of knowledge work has been a holy grail through at least the last 3 main tech cycles - remember Knowledge Engineering, Artificial intelligence etc - and so far has still not got very far in practice]

6. Engaging the next three billion digital citizens

As incomes rise in developing nations, their citizens are becoming wired, connected by mobile computing devices, particularly smartphones that will only increase in power and versatility. Although several emerging markets have experienced double-digit growth in Internet adoption, enormous growth potential remains: India’s digital penetration is only 10 percent and China’s is around 40 percent. Rising levels of connectivity will stimulate financial inclusion, local entrepreneurship, and enormous opportunities for business.

[The next 3 billion citizens are very uninteresting by Western advertising economics standards, but Western price Ads (or selling data to advertisers) is required to fund the Western Dotcom industry at its current cost levels. It will thus be interesting to see the business models used for the next 3 billion, because most of the current Ad based ones won't work with Western based infrastructures]

7. Charting experiences where digital meets physical

The borders of the digital and physical world have been blurring for many years as consumers learned to shop in virtual stores and to meet in virtual spaces. In those cases, the online world mirrors experiences of the physical world. Increasingly, we’re seeing an inversion as real-life activities, from shopping to factory work, become rich with digital information and as the mobile Internet and advances in natural user interfaces give the physical world digital characteristics.

[Remember 2nd Life? All this was promised, back then. It tanked, and nothing has changed that reduces the ikelihood of another failure just yet]

8. ‘Freeing’ your business model through Internet-inspired personalization and simplification

After nearly two decades of shopping, reading, watching, seeking information, and interacting on the Internet, customers expect services to be free, personalized, and easy to use without instructions. This ethos presents a challenge for business, since customers expect instant results, as well as superb and transparent customer service, for all interactions—from Web sites to brick-and-mortar stores. Fail to deliver, and competitors’ offerings are only an app download away.

[Today, most of your competitors are still no better than you, finding the right App is a nightmare, it's not a given it'll work on your device, and there is a lot of stuff we still won't buy online - see below. But this is probably one of the areas that is tightening up faster than some of the others]

9. Buying and selling as digital commerce leaps ahead

The rise of the mobile Internet and the evolution of core technologies that cut costs and vastly simplify the process of completing transactions online are reducing barriers to entry across a wide swath of economic activity. Amped-up technology platforms are enabling peer-to-peer commerce to replace activities traditionally carried out by companies and giving birth to new kinds of payment systems and monetization models.

[Except that there is a large swathe of things we don't want to buy without physically examining them first, and a larger swathe we won't buy if we don't trust the seller, service, payment process etc. We are happy in the main with buying low cost, low risk commodities online. For everything else, we still like face to face communication. Its changing, but slowly, and in fact we are starting to see online retailers look at bricks and mortar stores to increase their penetration. Very few payment systems are "new", most have been thought about for 20-odd years (just look at all the dotcom initiatives), typically the blockers have been technology costs, vested interests or lack of user takeup]

10. Transforming government, health care, and education

The private sector has a big stake in the successful transformation of government, health care, and education, which together account for a third of global GDP. They have lagged behind in productivity growth at least in part because they have been slow to adopt Web-based platforms, big-data analytics, and other IT innovations. Technology-enabled productivity growth could help reduce the cost burden while improving the quality of services and outcomes, as well as boosting long-term global-growth prospects.

[I've been doing Internet strategy for nigh on 20 years, this area is always just about to take off in a big way. However data privacy, the problem of getting the last 20% to use new services to realise the financial benefits and related worries about democratic representation, the huge discrepancy between who pays and who benefits from public money, and expensive past failures are all big barriers to major adoption. We see no compelling short term drivers for major changes. In the medium term the cost of providing for the elderly, and public services, will force changes.].

We are not saying these things won't come to pass, but we are saying it is going to happen far more slowly than this paper implies. This will mainly be a slow evolution, not revolution. To help you work out what is real, what is coming, and what is hot air we present the Broadsuff Modifiied Gartner Hype Curve, complete with the Perpetual Hype Re-Cycle for those topics that go up and down the curve time after time but never seem to create a new service.

*I've just invented it, but I'll bet it exists by 2015 as anyone who has worked with big datasets or simulaton models has seen this phenomenon. The behaviour of anchoring - seeing things you want to see that aren't there, or stopping when you see something you like and not testing for anti-patterns - is also already well documented

Amazon and Fan Fiction - the Rights stuff

nospam@example.com (Alan Patrick) — Wed, 22 May 2013 17:35:27 +0100

This is a very interesting development - TechCrunch

Kindle Worlds joins Kindle Singles and Kindle Serials as a way for authors to earn money from digital publishing, and the best part is that in this case you don’t even have to be all that creative – the idea is to let fans create stories around original properties from other authors, offering them up for purchase on the Kindle book store. Amazon then pays out royalties to both the original rights holder, as well as to the fan fiction author, with the author making around 35 percent of all net revenue for works over 10,000 words.

There’s also a new revenue model aimed at shorter works, which would be between 5,000 and 10,000 words and will typically sell for under a dollar. Under this scheme, the author’s cut will be a digital royalty of 20 percent.

Fanfic writers can sign up now at the official Amazon Kindle Worlds website, and the company expects to launch the Worlds storefront in June. There will be over 50 commissioned works included in the store at launch, Amazon says, and then it’ll be launching its self-serve submission platform for all authors to add their own completed works for consideration.

Kindle Worlds will initially just have the licenses to the Gossip Girl, Pretty Little Liars, and the Vampire Diaries book series. The Mary Sue asks all the questions that occurred to me, so allow me to praise by copying it:

This whole thing raises a slew of questions, among them how transparent the guidelines to determine what “reasonable” will be, whether Kindle Worlds will accept fics already published, how many people will be willing to pay for something that’s always been free, how much Amazon will charge, and if/how the availability of free fanfic will be affected. I’d also be curious to know how much the rights owners get compared to the writers’ 35%. I can see being cynical when it comes to the money aspect of this; giant corporations (the publisher of Gossip Girl, Pretty Little Liars, and Vampire Diaries is owned by Warner Bros.) will be making money off of the labor of their fans. That’s not a viewpoint I share, though, because that’s what happens anyway: Fans put thousands of hours of effort into creating fic, graphics, crafts, etc., expecting nothing in return other than the object of their fandom being good. The fact that some of those fans will now be able to get paid for their efforts if they so choose is amazing

Finding a way for aspiring authors to not just write fan fiction, but build a reputation by siting them in existing (well read and loved) worlds. This interests me because its an early sign of how content rights are being rationally restructured for the digital world. Easy to do, low cost - Ronald Coase would be nodding, saying "I told ya"

I just pray they never get the rights to Middle Earth....

Social Business Reimagined as a Coasian Construct

nospam@example.com (Alan Patrick) — Tue, 21 May 2013 23:51:00 +0100

Soundbite video from the Business Reimagined session last week, I used the term "Coasian construct" for Social Business, thought it may need a bit of background.

Firstly the Coasian bit - that is referring to Ronald Coase, a 1930's economist, and his thesis on transaction costs. Wikipedia sums it up perfectly:

The traditional economic theory of the time suggested that, because the market is "efficient" (that is, those who are best at providing each good or service most cheaply are already doing so), it should always be cheaper to contract out than to hire. [This was talking about the 1930's, not the 2010's by the way - not that you'd notice any great changes )

Coase noted, however, that there are a number of transaction costs to using the market; the cost of obtaining a good or service via the market is actually more than just the price of the good. Other costs, including search and information costs, bargaining costs, keeping trade secrets, and policing and enforcement costs, can all potentially add to the cost of procuring something via the market. This suggests that firms will arise when they can arrange to produce what they need internally and somehow avoid these costs.

There is a natural limit to what can be produced internally, however. Coase notices "decreasing returns to the entrepreneur function", including increasing overhead costs and increasing propensity for an overwhelmed manager to make mistakes in resource allocation. This is a countervailing cost to the use of the firm.

Coase argues that the size of a firm (as measured by how many contractual relations are "internal" to the firm and how many "external") is a result of finding an optimal balance between the competing tendencies of the costs outlined above. In general, making the firm larger will initially be advantageous, but the decreasing returns indicated above will eventually kick in, preventing the firm from growing indefinitely.

Other things being equal, a firm will tend to be larger:

the less the costs of organizing and the slower these costs rise with an increase in the transactions organized.
the less likely the entrepreneur is to make mistakes and the smaller the increase in mistakes with an increase in the transactions organized.
the greater the lowering (or the less the rise) in the supply price of factors of production to firms of larger size.

The first two costs will increase with the spatial distribution of the transactions organized and the dissimilarity of the transactions. This explains why firms tend to either be in different geographic locations or to perform different functions. Additionally, technology changes that mitigate the cost of organizing transactions across space will cause firms to be larger—the advent of the telephone and cheap air travel, for example, would be expected to increase the size of firms. On a related note the use of the internet and related modern information and communication technologies seem to lead to the existence of so-called virtual organizations.

I've put the last line in italics. To explain a bit of history, I started looking at the impact of information flows on business in the mid 1980's for my MSc, and that's when I came across Coase and transaction costs, and it became clear to me that the Internet (as it came to be called, it wasn't called that in 1985 when I started playing with "wide area networks") was going to have a massive impact on business transaction costs. Researching back, it became clear that the main impact of all communication revolutions in the past had been to reduce transaction costs, and once that became clear it was possible to predict the impact of the Internet far more closely, so for example the arrival of things like Amazon, eBay etc ("Silicon Soukhs", as we called them in the early 90's before the dotcoms existed) were entirely predictable.

Secondly, the Construct bit. Now we are seeing the next jump forward in Comms technology, Social Networking. We started looking at their impact on business seriously in 2005 when we founded Broadsight, as we could see what the rise of blogging, wikis etc, and what better webtools would do to Web 1.0 technologies like groupware (If anyone would like our classic 2005 paper, "Everyone is a communication business now" email me now ). It was clear that what they would do would be to have another major impact on another layer of transaction costs, that between individuals both in any enyerprise and in the value chain, ie it would go down to the person to person messaging level. Previously the major impacts here had been the telephone and email, this would take the costs down another order of magnitude. That was also why Twitter was so interesting when it emerged, as although we had great fun laughing at the "what I had for lunch" tweeters, the bigger point was that the transaction costs had reduced to the level you would say this over an electronic medium.

In essence, what this means is that with Social Media the whole enterprise can, (in theory - see below), start to resemble the buzz of conversation around a small company's workspace, even if it is larger, geographically, and even temporally dispersed. This buzz is the stuff that makes businesses hum, that ensures many little things "go right" and its why businesses that get above a certain size they start to become much less efficient ("decreasing returns to the entrepreneur function", including increasing overhead costs and increasing propensity for an overwhelmed manager to make mistakes in resource allocation, as Coase would have it). Furthermore, Social Media is continuing outside the enterprise, so in theory the buzz of your potential customers, suppliers, competitors and existing users are also all in the room with you. The business efficiencies that are theoretically possible from this are immense. Just imagine all the little inefficiencies you see in your company, and in those you transact with, and imagine a major reduction in that.

It would be a massive change in efficiency for any enterprise.

But, what still has to be controlled is data and message overload - that is what has killed email's usefulness, the lower transaction costs of writing and sending an email has not been matched with a similar reduction in the cost of reading and acting on the message - and it wil kill social media in an enterprise if it is not managed better than email. The term "Big Data" to me is an admission of failure, its occurring because we have not yet worked out how to target what we are looking for. The major risk in social business is drowning in data, and that will be the big challenge for social business systems.

Tumblr'ing Down

nospam@example.com (Alan Patrick) — Mon, 20 May 2013 13:56:48 +0100

Hot on the heels of the wisdom (or not) of acquiring Summly, Yahoo has now apparently also acquired Tumblr for $1.1bn. Tumblr, for those of you who don't know, is somewhere between a a blog and a microblog (a Mediblog), has revenues of $13m. So why pay $1.1 bn? The argument is the user numbers - Next Web:

With more than 300 million monthly unique visitors and 120,000 signups every day, Tumblr is one of the fastest-growing media networks in the world. Tumblr sees 900 posts per second (!) and 24 billion minutes spent on site each month.

So lets value it the old, boring way:

300 million x 12 = 3.6 billion unique visits (not necessarily different visitors of course) and 120,000 new ones a day x 360 days = c 43m new users per annum, the hope no doubt being that this will continue to grow like topsy. Lets assume that this growth gives c 1 billion users in 5 years, so we get a 50% increase PA on 50m new users, and assume we lose none, and that gives a Net Present Value of about $ 0.33 margin (at 15% IRR) to get to c $1.1 bn fully discounted free cash flow.

Or thereabouts.....we can also do it another way - by comparison:

Another Mediblog, Facebook, has an Average Revenue per user is about $1.25, and it's user base is about 1.1 billion and slowing. It is valued at $60bn in an open market. Applying Facebook's valuation to Tumblr today, with c 1/2 the user base gives us $60 bn x 1/2 user base x (0.03/1.25) ARPU = $0.75bn. Throw in a 33% uplift fir future optimism, and there's your $1.1bn

So you can believe on of two things:

(i) The Bubbletime will continue after Yahoo's acquisition, and the business will stay as good as Facebook is now, or

(ii) Yahoo's acquisition will cause a near immediate jump in ARPU from $0.03 to something like $1.25, it will better Facebook by extracting not C $0.01 profit per user but more like $0.33 profit per user per annum.

But of course you can - or Yahoo thinks so anyway, heck they even paid with cash, not shares!

Alternatively, one could take the opening line from Cockney Rebel's "Tumbling Down" anthem.

"Gee, but it's hard when one lowers one's guard to the vultures"

The refrain of which is Oh dear, look what they've done to the blues, blues, blues.....

The Evolution of Online Psychology

nospam@example.com (Alan Patrick) — Fri, 17 May 2013 10:51:45 +0100

Implicit relationships people seek from products (source: Simon White, Draftcb)

Last week I attended another event i think was very useful seminal in pushing Social Media forward, the ChinwagPsych event, which reviewed the emerging lessons from the emerging fields of digital psychology and anthropology. It was really good as well because the amount signal was high, and the amount of noise from snake oil, bullshit and hype was low. Here are some notes of mine from the event.

Nathalie Nahan, who wrote Webs of Influence, on Website design in the Social era.

- Trussst is what stops people buying online, as customer service/delicvery etc is unclear. Consumer reviews are good at creating trust, also earned media. Using recognised logos eg paypal also helps
- Often need to give people some sort of value upfront as unknown/ untested vendor (Freemium, 30 day trial etc)
- 30% of online shoppers and rising worried about privacy (about bloody time too)
- When they scanned websites to get missing f2f trust cues, they found yellow was the worst colour for trust, followed by red
- Websites need a very defined call to action button (Now this is really Web 1.0 old hat, but since Nathalie's talk, I've become really aware of how many sites still do not do this)
- Using the scarcity principle - "only X left in stock" - to create a sense of urgency, really works
- UK is culturally difficult to US, need to soften the hard sell -
- Tweet to unlock benefits seems to work well
- Humans focus on contrast, and on concrete stuff - eg money management website example "I will teach you to be rich" is succinct and to the point.
- Framing in terms of loss raher than gain is better, as Humans are more loss averse
- 10% of a population = tipping point, when people start to follow without thinking (Interesting, I've seen memetic algoritms that calculate that when say when cheating gets to about 5% of the population it then starts to spread rapidly)

Leigh Caldwell, Inon - Psychology of Pricing

- This is an emerging new field - cognitive economics
- Talked about improving profit via the "rule of 3" - Typically with 2 choices get a 70/30 split, 70% cheapest.
- Insert an expensive 3rd choice - 3 choices split 60/35/5 - fastest way to improve profit
- Concept of "anchoring" - set expectations of higher price first, people will choose next cheaper option
- Human brain treats spending money siilarly to pain, so give people something or offer a later payment
- If you make price complicated, befuddles customer, reduces ability to discern (I tried to ask Leigh why mobile Co's, Energy Co's etc do this, despete stromg evidence it puts off customers but ran out of question time)
- You must Price differentiate for different customer segments to maximise profits, but be prepared to lose some sales
- People will value what you price

Simon White, Draft Cb - Why are people NOT using all this stuff?

- Old models (persuasion) only lead to accidental success as 95% of decisions are instinctive, not rational/ logical
- People don't like change/new ideas, so try to rationalise to their past knowledge, and it seems to work "well enough" as new approaches haven't really had major impact yet
- New approaches are still complicated to use, no rules of thumb, no benchmarks
- Rapid Change, new "New things" every 6 months - overwhelming amount of data coming out and overwashing last new new thing
- "Institute of Decision Making" set up, linking Research to actual Application of these ideas

The new thing I got from Simon's talk was this:

- People buy things to achieve goals, they do not have a "relationship" with brands as such. And their real goals are the implicit goals, eg buy Mercedes to feel powerful
- 3 main implicit driver axes

(i) security
(ii) excitement (drives dopamine)
(iii) autonomy (drives testosterone)

- Intersections between the above 3 exist - then adventure, discipline, enjoyment (see my diagram at to of page)
- So move proposition to answer explicit and implicit goal. E.g. Oreo - explicit goal, treat child. Implicit goal - passing on your experience, childhood

Simon Hill, Wazoku - Open Innovation - A rereshingly free-from-hype talk about the reality of Open Innovation.

- technology is not the issue in Innovation
- people need to see the benefits,
- also must sort wheat from chaff
- need to put culture in place to force innovation
- Idea management is about process of generating, selecting and executing ideas, not about the 1 big idea
- Only about 5% of ideas will get executed

I then had to go to a client meeting, so missed a few talks and lunch (boo), returning in time to catch:

Steven Haggard & David Stillwell, Cambridge Predictive
I first heard about them awhile ago when they started to put together patterns of connected likes, on Facebook, they mentioned some typical insights e.g. people who like Terry Pratchett and computing tend to be introverted and Likes on Facebook can tell relationship status (I'll bet!!). They started to look at Business Applications, for CRM, trying to find predictive qualities of personality for honing the Marketing message; keywords, behaviour types etc. They mentioned as an example the Relentless fizzy "energy" drink going up against Red Bull etc:

- Its roots were in music roots so they started look at interconnections there
- looked at what energy drinkers like, by age, sex, etc etc on FB, and clustered the data into 200 like groups with interlinks,
- Saw that Relentless drinkers unique in Top Man U, Bad Diets, Classic heavy metal. So th
- Built Matrix of Energy drinks vs like groups, with probabilities on intersect.
- Overlay like groups with demography, and created a new new psychological profile layer above the "correlation" layer
- Relentless users are impulsive, neurotic etc, create narratives to fit
- Found that Music choice is a good proxy for personality index overall, so you can see eg Aerosmith and Journey hit different people
- Now working on Numerical score of musical affinity

Cat Jones, Unruly Media - Viral Video Chart

- built Unruly Sharerank algorithm by testing large no of hypotheses of what drives sharing vs actual sharing
- Myth #1 you can't predict sharing - you can, and quite accurately
- Myth #2 content must be funny - but you can do warmth or excitement as Main vector for eg
- 2 Video examples predicted - "3"s Pony doing a Fleetwood Mac Dance = funny and happy and warm; and Pepsi's Camaro = funny and nervous and excitement
- Camaro did as their aglorithms expected, Pony did much better - mainly due to an unpredicted factor, the horsemeat scandal at teh time
- Viral peak is usually day 2, on avearge 25% of all shares on 1st 3 days
- optimal time of video = c 3 min (We found the same for video clips when testing in 2007)

Milward Brown (Similar findings to Simon White, above)

- The only people who were shocked by behavioural economics were economists (rational man)
- $35bn Market prediction industry, only 2% modern tools - why?
- traditional tools are not that broken
- New methods are not that simple to use, and hard to transfer from specific project
- emotion is not everything, most people have very little resonance with most brands most of the time
- New methods give very different results but case law does not exist to calibarte them - very little published success.
- Also they are hard to understand

What was new is they have tested a number of these new approaches against Ads

- Most useful method they have found is facial response, map facial
happiness as Ad plays. Integrated with survey data from same people
- Shows key impacts across different cultures, and what people react to vs what they say they do
Can also check reactions at 2nd view etc.
- Online Behavioural Psych is more Nudge than Revolution

The next 2 talks were about Smart work, and were interesting in that they seem to apply a "psych" layer to the Quantified Self movement

Design by Day (Book on smart work) Nokia/Anthony Mayfield

- using mobile technology to improve productivity
- if you get 5 hours sleep a night you are no better than being drunk
- only a decade into ubiquitous connectivity, 3 years into smartphones, so started to look at what can be done
- use Design thinking to structure workflows, design Prototype Days to test what works/doesn't work, timeline a bad day, see what you can learn.
- Also looked at older ideas that neuroscience backs up
- Thinking is expensive, glucose is measurable - habit is lowest energy levels
- work day doesn't match how our brain works. We only have 3 - 4 hours a day at peak performance, yet we often waste it on doing emails and meetings
- Benjamin Franklin was a major designer of his day
- Need to close off little worries
- Know when to switch off, scheduling time for email and Twitter etc - McKinsey report showed multitasking is less efficient on all tasks

Prof Karen Pine - "Do something different"

- Use technology to remind you to do something different that is good for you, force behavioural flexibility
- people tend to use some behaviours too much, drives personality (locked in habits)
- brains like to save energy, repetition drives strong pathways. But this default is not always right way in every situation. Success can over- embed behaviours.
- If there is a gap between our behaviours and what is needed, it drives stress, eg: strong extrovert may not have introversion capability, creates stress
- Most effective way of changing people is not change what they know but change what they do, as otherwise they flip back to what they know how to do
- People develop by doing new things, trying to do same stuff in new ways

Benjamin Ellis, Social Optic

- how to use technology to change larger groups - uses social data to drive changes in organisations
- Amount of data in businesses has grown hugely, used to have data scarcity, now glut in flow and storage
- Unstructured data is the major challenge, we now store it and work out how to structure it later (I agree with Benjamin here, this is the major issue wiyh a lot of Corporate data)
- Old way of marketing - ask people, but they tend to give acceptable answers - different Co cultures drive different mindsets - measure attitudes, not behaviours
- New way - scrape social media behavioural data - tells are not only what people say, do - but how they do it eg how long they take - aim is to predict behaviours
- A lot of social network data is not behavioural, it's managed attitudinally
- In businesses using social media proxy data, main lesson is the shape of the graph
- Language used is an interesting tell, in the business social media, you can tell a lot of things from it, eg Sentiment
- Look at where people use common language, shows the social influence graph
- But can wind up with "beautiful mind" syndrome, seeing many many potential patterns
- Also need to factor out correlation, to only get causation - its very hard
- People also follow biases when given Big Data, anchor on the 1st thing they see, confirmation bias and bandwagon effect can take over
- Chris Anderson saying sheer volume would obviate theory - wrong. Nate Silver book better
- Game mechanics is using psychology to influence behaviours

Daniel Bennett & Marina Clement, Ogilvy Change: Case study - the missing £2

- Rory Sutherland started behavioural economic arm, heavy usage of outside academics. Case study was that people are stopping buying newspapers, so how do we sell them again. Approaches were
- 3 option choice architecture with dummy option (like Economist used)
- differentiate between tablet, PC and smartphone offerings
- Nudge 1 choice sleight of hand in use of deals (see Economist above)
- Nudge 2 choice overload
- Nudge 3 superiority bias - "ultimate" pack
- Nudge 4 - create easy to choose default

Those 4 "nudges" hit 257% ROI, I asked which ones had the most impact, they said they didn't know. I find that hard to believe

Fascinating and worrying in equal measures.....

Applying Social Media to Business

nospam@example.com (Alan Patrick) — Thu, 09 May 2013 22:30:37 +0100

Broadsight Simplified Value Creation Flowchart

We have been asked to be on the Microsoft Business Re-Imagined panel on the 15th of May, looking at the potential impact of Social Media in business. We've been consulting to and building social media systems for clients since 2005, and it seemed like this was a good time to boil down our experience down. Above is a simple value creation flowchart, and we look at how, in our experience, social media can impact value creation. Stripped to it's bare bones, a business creates sustainable value by increasing revenues and/or reducing costs. Social Media is a new set of technologies that can help in a number of ways.

Revenue Increase

Essentially, there are only 2 ways of increasing revenue, which is sell more stuff, or increase the price at current volumes. In my view, Social Media is more powerful (with current technology, anyway) in dealing with Consumer sales, rather than Business sales, as the main cost in a consumer sale is reaching the large mass of disparate consumers. With business sales the target market is much easier to identify, and there are less customers to contact so current approaches work well.

Sell More - Social Media is a potentially very powerful tool to generate new demand - to find new customers, to bring them to the enterprise, create trust, ease their purchasing journey, and help optimise website design to maximise sales throughput. Taking it to greater extremes, Social Media data can be used to start to psychologically profile your customer base to understand what sort of people may also be your customers. It is also a very powerful tool to optimise the product - to discover the features or configurations that customers really value, and you can identify what competitors do well and include those in your product. But it is not a substitute for existing methods today, it is an adjunct, it can't be the only tool. It is not an appropriate tool for overt "hard sell" marketing, for example.

Increase ARPU (Average Revenue Per Unit) - Social Media has two main functions here - to aid in communicating the value proposition of the product, and to improve the perceived value vs competitive offerings. Communicating the value is fairly straightforward, as it serves as an extension of traditional channels, with the added benefit is that it has a strong feedback loop so it is quickly possible to see what is working and what is not. It is also probably better at communicating implicit values than more traditional media. Social Media can also help to improve the pricing point by optimising product differentiation amomg dofferent customer groups, and find pricing points for these different customer groups. That feedback loop, and the data it drives, can be used to to find attractive product combinations, and to optimise the website design to maximise value per sale. Taking it to greater extremes, Social Media data can be used to start to psychologically profile your customer base to understand their hot buttons better

Cost Decrease

Again, essentially there are only 2 ways to reduce costs - reduce operating expenses (Opex) and reduce Churn (Customer defection). I will ignore Capital expenditure (Capex) for now. Social Media in my view is as useful in both consumer and business facing enterprises, as it's power is about increasing productivity and effectiveness in reducing costs

Reducing Opex - In most companies, there are two major cost areas - raw materials and labour, the sum of them is typically the "80/20" of the cost pie (companies with near zero raw material costs tend to be professional service businesess, with very high labour costs). So far in or experience the two major Social media impacts are from:

(i) getting better market information (both pre and after sales) from customer to company, allowing the company to both reduce costs or lead times of raw material while not reducing value, and being able to better place its human resources where it really matters - it can work like a real time value engineering approach. It's not just useful for line operations, social media can be used to influence better design and innovation, and can be used to increase "brain cycle cpacity" by tapping into customers and the overall milieu without having to employ it

(ii) getting a better flow of information between people in the company, so co-ordination is better (less balls are dropped because A didn't know what B knew about customer X) and spped of reaction is faster.

The third cost element, overheads, is interesting. I have some thoughts about using Social Media transaction data to better allocate some overheads, but I don't have a fully formed set of ideas and I haven't seen it put into practice anywhere yet.

Reducing Churn/Increasing Retention

Churn is often not well understood, as many businesses are not aware of the huge differences in cost between retaining an existing customer, and finding a new one. There is often an order of magnitude difference. Thus reducing churn can have a massive impact in cost reduction and revenue increase Social Media can be used to aid customer service, to serve as an early warning for customer problems, to find out what people really value in post sales service, and to improve the product lifecycle. Social media also means bad service is more likely to leave a company's reputation punished, which can also impact sales, as customers typically research online before buying. In saturated industries, having a better churn than the competition can radically alter the market share and strategic positions within a few business cycles.

A brief word on Capex costs - if Social media is helping a busines to make better use of existing assets, in theory it will slow down future Capex requirements - but Social Media technologies do have Opex and Capex costs of their own, and these are typically incurred early up, while the benefits are then gained over a series of cycles. Which brings us to the dreaded Return on Investment (RoI) question.

Return on Investment

For any new technology, ROI is hard to work out, as few case studies exist to allow calibration of cost and benefit. If history is any guide, new and risky ideas are typically implemented first where forces are greater than a pure ROI worry:

- Piloting: trying out the New new thing in some areas of the business, somethimes structurally, often though by "Intrapreneurs" who do it locally out of passion, or seeking promotion etc.
- Pressure: A company realises it will not succeed doing "Business as Usual", and has to do "Business as Different". Recessions are for this reason more likely times to see new ideas implemented
- Politics - needing to be seen to be "with it", intra-divisional rivalry etc - all these can drive early day projects

But eventually, for Social Media to take off and scale, believable ROI needs to exist. We are dubious of some of the various "Returns on" currently touted for Social Media, as it is hard to tell which are valid proxies for hard to measure benefits, and which are just Snake Oil. Our test is that if a proxy measure cannot be linked to an underpinning economic benefit logically, its more likely than not to be snake oil.

Other Impacts

So far we have seen 3 meta-impacts of Social Media occurring

Firstly, there is a synergy when multiple of the above areas are pursued. The marketing listening systems can influence the customer satisfaction and product design systems, the customer satisfaction system can improve the marketing message, better staff co-ordination can improve customer service and salesforce knowledge and thus effectiveness. In general as more information flows in a business, better decisions are made throughout the business. But its not a given - if organisation culture is not open, if individual reward structures do not encourage sharing, if management use the new dataflows to further expolit staff, these systems can go nowhere, even potentially accelerate problems as they are just better ways of making sure all the sh*t hits the fan.

Secondly, the same influencing ability that works on attracting potential customers can influence sentiment, and thus share price. But this has always been a temporary game in the past, and its unlikley Social media will be any better at selling the sizzle if there is ultimately no beef.

Thirdly, expanding on this, a basket case will be exposed far more quickly - disaffercted customers, employees, investors etc will make their feelings known. In teh old days, they sued to say one disgruntled customer would on average tell 7 people. Now they will tell tens, hundreds, thousands, will write on review sites, will push negative pages higher in the search rankings than "official" company narraticves, etc etc.

Beware the Snake Oil

There are too many Social Media snake oil merchants promising miraculous cures today. We talk about what Social Media can do above. This is what it cannot do:

(i) Replace the modus operandi of most businesses. Social Media is essentially a communications medium, like telephony or IT. Telephony made a major difference to the cost of selling retail insurance for example, but it did not replace insurance. Web sales has restructured the book industry, for example - but not replaced it.

(ii) Replace existing operating techniques - marketing and sales techniques, service techniques, working practices etc - or not straight away, anyway. This is for 2 reasons. Firstly, existing techniques exist because they still work a lot of the time, and Social Media is not an appropriate complete replacement, but is an adjunct to improve them in most cases. Now it may come to pass that eventually Social Media might replace, say, telephone call centres - it's plausiblee - but than a bunch of other things need to change too so it will more likely be a phased change, not a big bang. Secondly, history tells us that any new enabling technique actually doesn't replace the old, it usually just slowly supercedes it in the pecking order. The revolution will (in the main) be evolutionary.

(iii) Massively replace employees with other peoples' brain cycles - "crowd -X'ing" is overblown, because most businesses do not win the day on one-off surge efforts, but on day to day execution over many cycles, and to do that you need well trained and motivated teams, not an ever-shifting cast of part-time volunteers. Adhocracy is a great way of getting Ad hoc one-offs done, and we are all for it in its place - but its not a great way of delivering reliable, predictable products and services day in day out.

(iv) Be the Silver Bullet that saves the day. It will need to integrate into a number of other existing systems in any enterprise. Social Media is a new layer of enabling communication technology, that makes existing practices and processes more productive, or effective, or both. The degree of just what, where and how it is apporopriate will vary by company and industry.

(v) Ensure that This Time Things Will Be Different, save the Planet, bring Universal peace and Love, etc etc. Humans are humans, with the same foibles as they have ever had. Social media is not going to bring about a Business Utopia, where people give warm personal service all the time, all goods are maximum quality at minimum price, and cheaters never prosper - but because it lowers transaction costs, all those things will become a bit easier to police - and thus new ways of cheating, skiving and flogging lemons will emerge.

If we had to give our sanguine view, Social Media is the sort of cluster of technologies that today can give low % increases to all those areas we cover above - but once you sum up say a 5% increase in sales volume, a 5% increase in ARPU, a 5% reduction in OPEX and 5% reduction in churn, you wind up with a shift from (ay) 5% margin to near 30% margin. That is life changing for any business, but it has to be done within the current business systems to get the full impact. The medium may be the message, but it isn't the modus operandi*.

The Downside Risks

As with any powerful new technology, used improperly it can blow up in ones face. There are two main dangers that have emerged so far

- Leaky Ships: Sensitive information will leak more easily, so there does need to be more attention paid to keeping things tight where required. This is of course in direct opposition to the need to be open, and finessing the systems that handle commercially important data is still an emerging area and in our view is still a limit to Social Media reaching its theoretical potential.

- The Viral Faux Pas: The inappropriate and inopportune tweet that goes viral and pours opprobrium on the company is a frequent enough occurrence to be a Social Media standard trope. Apart from blaming it on the Intern, companies do need to be careful and have checks on their output, and damage limitation measures in place

The devil in doing all of this is of course in the details. So, we hope to see you on the 15th.....

(*With the exception of wholly new businesses that are Social Media businesses, of course - but that is the subject of another post)