Tuesday, February 25. 2014
Impact of mathematical techniques on operations, by industry - McKinsey
McKinsey has discovered you can use Operations Research (or Decision Maths as it is known these days) mathematical techniques to analyse and optimise manufacturing operations - McKinsey Insights:
The application of larger data sets, faster computational power, and more advanced analytic techniques is spurring progress on a range of lean-management priorities. Sophisticated modeling can help to identify waste, for example, thus empowering workers and opening up new frontiers where lean problem solving can support continuous improvement. Powerful data-driven analytics also can help to solve previously unsolvable (and even unknown) problems that undermine efficiency in complex manufacturing environments: hidden bottlenecks, operational rigidities, and areas of excessive variability. Similarly, the power of data to support improvement efforts in related areas, such as quality and production planning, is growing as companies get better at storing, sharing, integrating, and understanding their data more quickly and easily.
Not only that, but you can apply Lean operating techniques in manufacturing companies too:
Nonetheless, to get the most from data-fueled lean production, companies have to adjust their traditional approach to kaizen (the philosophy of continuous improvement). In our experience, many find it useful to set up special data-optimization labs or cells within their existing operations units. This approach typically requires forming a small team of econometrics specialists, operations-research experts, and statisticians familiar with the appropriate tools. By connecting these analytics experts with their frontline colleagues, companies can begin to identify opportunities for improvement projects that will both increase performance and help operators learn to apply their lean problem-solving skills in new ways.
Amazing stuff....except its very, very old news. Monte Carlo simulations and capacity planning algorithms have been around for decades, a lot of it even pre-dates WW2. Value analysis started at 3M in the 1960's. Richard Schonberger wrote the groundbreaking Japanese Manufacturing Techniques in 1982 (I still have my copy) and he was merely Westernising something the Japanese had been doing for 2 decades by then. And then I saw this, which really made me smile wryly:
Similarly, a leading steel producer used advanced analytics to identify and capture margin-improvement opportunities worth more than $200 million a year across its production value chain. This result is noteworthy because the company already had a 15-year history of deploying lean approaches and had recently won an award for quality and process excellence. The steelmaker began with a Monte Carlo simulation, widely used in biology, computational physics, engineering, finance, and insurance to model ranges of possible outcomes and their probabilities
The wry smile was because I did much the same, in 1994-5, for a steelmaker, using some of these exact same techniques - while I was consulting at McKinsey to boot. I have the obligatory picture of big rolling mills from a grateful client, and the prize I won in the McKinsey internal "Practice Olympics" to prove it In fact I'd bet the McKinsey Quarterly in the 1970's, 80's and 90's will be full of analyses like this one. There truly is nothing new under the sun.
But with New Improved Big Data it can all be rebadged bright and new....except it doesn't work this way. There was a shedload of Big Data in the Old Days too (shop floor data capture techniques underpin most of the Internet of Things, and did you know some of the first broadband networks in the world went in at manufacturers in the 1980's). Manufacturing has always had a lot of data, and Big Manufacturers bought Big Iron to process Big Datasets then too (except it was called data with a small "d" then). The Monte Carlo methods, or N jobs on M machines Optimisation (for examples) are still the same algorithms they were in the 1930's and 50's.
And you know what - you just cannot simulate the minute operation laden details of a shop floor or logistics network reliably. No matter how big your dataset, or your computers, or your machine tool onboard intelligence, there is just too much variability. Which is why the Just In Time/Lean movement came about as the better approach - the aim was to simplify the problem, rather than hit it with huge algorithm models and simulations so complex no one fully understood what they were doing anymore (just ask the banks what happens going down that route) - the aim of JiT/Lean was to actually reduce the problem variability, to get back to Small Data if you like.
And you know what else - despite the analytical miracles I and many others performed in the day, despite the extraordinary efforts by managements and workers, so many of those steel mills (and clothing companies, and manufacturers of a million other widgets) moved East. There is only so much you can do against cheap labour, national subsidies and guaranteed government contracts.
And that brings me to something else in the story, which is what is really going on here I suspect - its not Big Data, its Big Economics:
Sure, its partly about raw material prices changing - when they are too high to buy or too low to sell you really have to be efficient at manufacturing. But when you are getting to this level of number crunching, after 20 years of Lean projects, in my experience it's because the endgame is appearing on the horizon, its a last of the summer wine story, the end of an S curve. Interestingly, it seems like all the McKinsey consultants and the project were in India, and Eastern labour costs are rising, as is oil for those long ship rides back to the European and US markets, so much so in in fact that there is an increasing trend in re-shoring, as production is coming back to the US and EU. Big picture, the low cost Eastern windfall is ending, and you have to start getting much smarter again about the actual manufacturing process. You can get benefits from doing it right with Big Iron and Big Algorithms, no doubt - but this sounds like back to the future....I suspect they are now using bigger and bigger number crunching to eke the last 20% of improvements from the various kaizen projects ongoing, trying to keep the factories in situ as the Big Economics shift yet again.
And you didn't need Big Data to tell you that....
(Hat tip to my colleagues at the Agile Elephant for the link)
Monday, February 24. 2014
The (non) regulatory annual Bitcoin crash
News today that Mt Gox may well have been turned over - Forbes:
This was always going to happen, as we've pointed out before, and there is no restitution - no one is insuring bitcoin holders against losses. And, just as predictably, post crash there will be regulation:
It's probably going to happen again before that though, as Bitcoin's decentralisation and lack of oversight is both its strength and Achilles heel
Thursday, February 20. 2014
News just in that Facebook has bought WhatsApp for $SillyMoney - $19.6bn - TechCrunch:
With 450 million monthly users and a million more signing up each day, WhatsApp was just too far ahead in the international mobile messaging race for Facebook to catch up, as you can see in the chart above we made last year. Facebook either had to surrender the linchpin to mobile social networking abroad, or pony up and acquire WhatsApp before it got any bigger. It chose the latter.
Facebook couldn't afford not to have it, if someone else had bought it that would have made a direct attack on Facebook's chosen strategic way out of its own declining user engagement, i.e. mobile applications and messaging. Facebook is absolutely determined not to be overtaken by the "next wave" Social Networks. But its own "new wave" systems were kludgy, so the price of not having your lunch eaten in 2016 is c $20bn in 2014. There is a bit of irony in that WhatsApp strongly do not believe in advertisng, their founder once saying that:
"There's nothing more personal to you than communicating with friends and family, and interrupting that with advertising is not the right solution,"
Clearly $19bn is a mind changing amount
What this really shows is that the days of Facebook's organic growth are over, and from now on in they are going to have to acquire revenue, and thats a very expensive way of doing it at their size if you are continually needing to buy the guys who will eat your lunch. You had to believe a lot to believe Facebook's valuation - that just got harder. But its the Bubbletime, so all will be good - for a while
Update 1 - According to El Reg, Whatsapp does indulge in datascraping of a user's address book, so that does make it an interesting prospect for added value datamining.
Update 2 - Azeem Azhar of PeerIndex has a smart bit on analysis - if WhatsApp had stayed independent it still would have destroyed Facebook's mobile story, which really outlines why Facebook had to act to stop itself being eaten for lunch:
Wednesday, February 19. 2014
From the BBC, a report on a series of universities trying to build a system that can counter social media borne rumours, lies lies and gossip. The Pheme (maneda fter the Greek goddess of Gossip) is a collaboration between five universities — Sheffield, Warwick, King's College London, Saarland in Germany and MODUL University Vienna — and four companies: ATOS in Spain, iHub in Kenya, Ontotext in Bulgaria and swissinfo.ch, led by Kalina Bontcheva of from the University of Sheffield. Pheme will classify online rumours into four types:
Apparently different types of digital disngenuity leave their own type of digital footprints and can be recognized. The system will also look at the accounts spreading it and look for bots. Idea is then to search for information that is true from known sources and re-seed the stream of the original falsehood followed with "the truth". It will be ready late 2015 apparently.
The obvious flaw is if it can detect falsehoods, any half decent falsehood spreading system can detect it and re-seed the same trails. The other sad flaw is many people will rather believe a convenient lie than an uncomfortable truth. The war for the truth is about to be fought in the cyber-memespace to an unprecented degree - I wonder if there wll be a new subscience of memetics, called "phemetics".
Google Glasses, 2020 Vision
Google has published a series of "Don'ts" for Google Glass aficionadoes to prevent them being Glassholes - not because of the danger of nerds being disrepected, but because its bad for business - Don't #4:
And Do #2:
Ask for permission. Standing alone in the corner of a room staring at people while recording them through Glass is not going to win you any friends (see Don’ts #4). The Glass camera function is no different from a cell phone so behave as you would with your phone and ask permission before taking photos or videos of others.
But even this still gives huge permission latitude. Most people would find it creepy if someone was sitting in their social situation with someone recording everything, even if there was not a formal "no camera/mike/recording" sign or no real permision required. As TechCrunch notes:
Google’s challenge is not building the Glass platform, but training the general public to welcome Glass wearers into society. Glass’s future rests largely on the public’s acceptance of the technology. If, like Bluetooth headsets, it’s deemed nerdy or, worse, if Glass is lumped in with the NSA privacy scandle, the technology will be an also-ran. A lot is riding on Google Glass Explorers.
The problem is many "Explorers" - being tech nerds - have the social intelligence of...well, tech nerds - the result is increasing unease with Glass, and the emergence of No Glasses Allowed and similar campaigns' hence the publication of the advice. They'd probably also be well advised to give them to more socially aware people rather than the sort of geeks who will pay $1,500 for the privilege of testing them.
But if you think this is creepy, just roll this forward a few generations of Moore's law when you can wear a much less obvious recording device, and it can access a wealth of Big Data Crunching and layer some form of augmented reality over your vision. Here's a scenario - you can walk into a room in 2020, scan all the faces, and do a search of all the data held against those people from a whole range of sources - on Google, from their social networks, from open data given away, and, for a price, from hacked data apps that give you that little bit more. The picture at the top of this post imagines that - imagine a cocktail party where you could see all the dirt on everyone after a guick glass and google.
Creepy? You bet. Impossible? You're fooling yourself - all those in the picture have happened or may soon do so:
The other 2 cases are hypothetical, but could come from Government data already in the frame for being opened up (though I note today we have another 6 months grace for medical records).
Playing Charades at parties will never be the same again....
Tuesday, February 18. 2014
A Salutary Lesson - Zynga Shares (Source: Yahoo Finance)
News out today that King, manufacturers of the current hot mobile game Candy Crush, have filed for a $5.5bn IPO. We were asked what do we think?
- Based on a growth from a few $100m revenue in 2012 to $1.9bn in 2013, and real profits of c $570bn, a valuation of $5.5bn is tame by Dot Com 2.0 standards, but....
They may be able to replace Candy Crush with a "next hit" but the Gaming ecosystem is littered with the bones of companies that didn't The music industry's stage is littered with the bones of 1 hit wonderbands. The movie industry...well, lets see what Candy Crush 2 looks like.
Sunday, February 16. 2014
Kickstarter was hacked - The Verge:
Hackers breached Kickstarter's defenses and stole the information of an unspecified number of customers, the company disclosed today. The company learned of the breach on Wednesday from law enforcement officials, and quickly resolved the breach, Kickstarter said today. It did not disclose how the breach occurred.
Where there's bank account numbers, there will be hackers. Good news is they admitted it, many companies don't.
The issue for any user of any Web service is that the more companies you put your data with, the higher the probability that it will be hacked and identities pilfered. So what to do? A good plan is to leave the minimum data with any company, but companies seem to want more and more data that is largely irrelevant. In the absence of a "minimum data law" or a trusted 3rd party service, or obfuscation services, we think using proxy data wherever you can (ie data that leads to something else, that only then leads to you -eg an email account that is not "you" for example) is the only real option for Joe and Jo Average right now - if anything its going to get worse (see our post on the Dark Side of Open Data - of course this applies to all data, not just Open data)
I can imagine a world emerging where non-digital data is actually valued more than on-line data again, and where private or "Word Of Mouth" networks (WOMNets) make a comeback as the tradeoff between convenience and security shifts.
Bertrand Duperrin's summary presentation
I've written up the 2 days of case studies from the Enterprise 2.0 Summit I attended last week in Paris, they are over here - Day 1 and Day 2 - on the Agile Elephant blog. There are a number of common threads emerging from the studies. Being lazy, I've copied Emanuele Quintarelli's list to start with, based on his study (as they concur with my analysis), I've added my thoughts in italics:
- The project is explicitly supported and sponsored by the top management (70% vs 34% for laggards). Long lasting processes, technology and process change should be somewhat mandated by the formal organization. Informal projects are easy to start but need formal acceptance to embed themselves in the enterprise
Other stand-out observations so far from the case studies:
- Any system needs time to embed/mature/settle in (words varied, but the concept was the same) before it becomes stable and self sustaining, you can't "make a baby in 1 month with 9 mothers" as it were.
Over the 2 days I did sense a departure of what the case studies were showing vs. what some of the Social Business theorists were espousing, in general the case studies showed that pragmatism and evolutionary development (what Dachis' Dion Hinchliffe called "Sustainable Transformation") was the order of the day vs more revolutionary/dramatic transformational approaches.
Update - I have also put up Bertrand Duperrin's slides (at top) now that they have been posted up, it was a very good talk on the subject as well
Saturday, February 8. 2014
Eddie Izzard de Latine
Vel potest esse solebas, nunc demum urget ab Anglis translate it lorem fermentum. Post haec audivi quasi. Et omnia, quæ sunt Cæsaris Caesar reddens. Quod ultimum test, nimirum, est ut in ipsum interprentatur interprentatorum:
Non ne forte....
Hmmm...Dies ne des opus
Friday, February 7. 2014
Personal, Private, Portable - and baked into the infrastructure... not a Pi in the (Cloudy) Sky idea anymore
Worries about how private data in the cloud is has set entrepreneurial minds a-spinning - TechCrunch:
The article is about a hardware based private cloud system startup - a personal, encrypted cloud device that plugs into your home Internet via Ethernet and provides cloud storage for your digital content without having to go through any third party intermediaries, and with support for unlimited expansion of said storage — either by chaining together multiple Pixeoms, or by plugging in additional USB hard drives.
I got quite excited about this. Private Personal Clouds are not that new, but Portable ones are much more interesting. Just as big servers in the office was a passing through phase, big servers in the cloud will be too, and at some point where Moore's Law trumped Zuckerberg's Law, the Portable Private Personal Cloud was bound to appear. In fact we even did a high level design* for one for some other people a year or so ago, and at Broadstuff Towers Mr Short has been known to get very excited about getting his hands on Raspberry Pi for this sort off thing (and getting his hands on the little computer Pi as well...) so i guess there will be more of these devices to come.
And of course, we can't end this post without the inevitable reference to an Internet (full) Of (these) Things, and what that will mean for Fridges everywhere**
*The next step up from the paper serviette doodle.
More Broad Stuff
Poll of the Week
Will Augmented reality just be a flash in the pan?
Creative Commons Licence
Original content in this work is licensed under a Creative Commons License