8. • Fully automated
• 24/7
• High volume
• You can’t hire enough people to check the content
• You can fix some using automation
• Even then, you still can’t fix all of the errors
• The labels don’t think it’s their problem
• It ruins the user experience
• Results in loss of sales
• Increase headcount requirements and cost
• Makes it very hard to engage users, you have to work
10 X as hard just to stay afloat.
9.
10. He who loves practice without theory is like the sailor
who boards ship without rudder and compass and never
knows where he may cast.
(Leonardo da Vinci)
Or, you have to have analytics in place in order to form
your strategies.
(Me)
16. 1. Microphone audio
2. Cell Radio location
(100m-10km)
3. Bluetooth location
(10-50m)
4. WiFi location (25-
100m)
5. GPS location (3-
10m)
6. Camera video
7. Light light
8. Accelerometer
motion/force
9. Gyroscope angular
motion
10. Temperature
temperature
11. Pressure altitude
12. Carbon Monoxide
air quality
SENSORS
17. Locations automatically entered into your
calendar.
1. Where is A, B
2. Can you work out how they get from
one to the other?
3. What else do you know from this?
18. 1. Identify a piece of music
2. Identify similar sounding music
3. BPM, Key, Vocalist Gender, Mode, Time sig
DSP
19.
20. 1. Open phone microphone remotely 2. Identify car from engine note
3. I’ve now got a demographic profile….
22. The future?
For People
• Predicting things based on your context.
• Making your life MUCH easier, less input, lots more relevant output.
• Adaptive, self learning experiences & technology.
• More control over the actual value of your data.
• Greater transparency of government and how it operates
For Business
• Predicting outcomes, including revenue, with real accuracy.
• Insight about competitors at prices you can afford and that levels the playing field.
• Lots of companies who sell but don’t analyse data will go to the wall.
• Some sectors will go through massive changes, transforming end to end operations,
improving efficiency, reducing costs and risk. (There will be tears…).
• Greater transparency of how businesses operate (more tears).
The back end systems supporting the lifecycle.
There is no difference in the process when compared to physical SC management, just the technologies you use to build the various functions
However, the digital world has a hugely error prone ecosystem simply because so much of the data is wrong.
The physical world (let’s say a supermarket) simply would not accept this situation, and they don’t. They impose standards on their suppliers to better manage stock and reduce operating costs.
However, global licensing practice means that doesn’t work. Master from US to Germany, data input without the e acute etc
If the data is wrong - you could not build an order picking system that you could trust.
No ‘just in time’
No efficient warehouse storage
No efficient way of stocking the shelves without having scores of people do it.
In fact they are really advanced in this field and it makes a BIG difference.
Now let’s look at what happens in digital music, in the old days a load of CDs would go to the store, and people passionate about music would lovingly classify it and put it in the right place.
As a user, I just made the decision not to use your service. If I don’t know what to search for, I can’t search, so if I can’t browse accurately there’s nothing here for me.
As me, I have to come up with ways of dealing with this (and guess how low on the priority list it is for the business)
(Beenie Man as classical, Mozart as Reggae, you can almost see what happened there as this feed was from the same supplier).
As the search and content management systems development manager, I’m at a loss what to do except come up with ways of hiding this or spend a lot of time cleaning the data and work out ways of doing it in bulk. (answer, cleaning, you can be smart with how you populate the store, just like in real world retail)
Taxonomy is key, as are tools to support the management if it, 200 new genres a week, so spurious it makes you laugh, but you have to map them to your version of the truth.
What people are passionate about doing as their job is not necessarily what needs to be done.
You need experts in the domain and good proprietary and bought in tools
You need business buy in above all else because it will cost money. Make the business case!!
Analytics
you have to have analytics in place and know what questions you want answered before you do it
to do this you need data champion and to educate everyone including (especially) the boss.
By the way, the average american listens to 4 hours of music a day. Probably the same as the UK, and so it’s an activity to watch as it happens more frequently than shopping, travelling, sex, in fact apart from ablutions and sleep its almost the only constant. That’s why understanding how people consume data and what it means is interesting to so many marketers.
You have to have analytics in place and know what questions you want answered before you do it
to do this you need data champion and to educate everyone including (especially) the boss.
And now, with social media, SEO, PPC, it’s about the convergence of old and new architectures. However what’s important is getting it right
Insight tells you where to steer the business
Predictive analytics won’t just tell you what product to show a user next using a recommendations engine, it will help your component buying teams save money.
It helps with fraud prevention (santander man and his 6000 fraud indicators)
Components and precious metals for 100’s of millions of phones
Predicting how much ‘air’ is in the buying process saved this much money alone.
Social means analysing sentiment about the brand or product releases
It turns customer care into a highly responsive relationship when customers complain or contact your @handle
Profile of hardcore dance music listeners (what can we infer?)
Who listens to more wee small hours music than anyone else in europe?
It’s our good friends the swedes, academics go mad when they get hold of our data…
By the way the release date of music pegs your aproximate age, it’s simply how radio programmer s work when they try to attract their target demographic – which is what they HAVE to do under the terms of their broadcast license.
The country of origin of the artists you listen to tells us where you are from.
The mix of artists tells us a lot about your possible gender…
Marketeers typically want age and gender. Job done.
Context/content, fly to Finland get news about strike. Lots of sensors tells me this. GPS, wifi, IP, operator (roaming) so if I go to finland, show me news about the ipending baggage handlers strike at helsinki airport. Ir’s surprising what you can infer…
Guessing game…
Digital signal processing
1. Shazam
2. Recommend me more music (and change the 80/20 rule)
3. Keep me listening to stuff of the same mood and tempo.
But you can do so much more….
Nice car, what sort of people own one?
No, so if we know someone is driving one we have part of their demographic profile
Same applies to other things like locations (although restaurants are hard…)
Quividi – french company
The data is out there, you will be found out. Better to get ahead by preparing for the new era of accountabilty that will be imposed on people by big data and analytics
One word of advice, use professionals, it will pay off.