Disqus - Latest Comments for mediaczar

Re: Three laws of technology hype.

mediaczar — Fri, 21 Apr 2017 08:56:06 -0000

Corollary: Everyone believes that organisations other than their own are better able to plan, agree, and implement new technologies. Only after they've changed organisation several times do they begin to see a pattern emerging.

Re: Optimal Diet Type

mediaczar — Wed, 20 Jul 2016 05:30:56 -0000

I'm rather with Lyle here. I'm a fairly nerdy marketing type, so I'm aware of two different kinds of "personalisation." Loosely these are "bespoke" (everything is designed around the customer) and "made to measure" (a pre-existing model is adapted to fit the customer).

What you're offering is -- really -- neither. It's a "selection."

It's a good deal of money to pay for a diet plan that doesn't take into account:

1) My goals
2) My weight
3) My TDEE

I mean, I could be a 95kg male who's exercising regularly, and I'd get the SAME diet plan as a 55kg woman who's doing no exercise at all.

Also -- It doesn't give me macro goals (what constitutes "low carb", here?)

You're selling science, but delivering something fairly fluffy. I think

Re: Benedict Evans

mediaczar — Wed, 22 Jan 2014 12:11:24 -0000

Congratulations, @Benedict Evans. It's a fabulous opportunity.

Re: The Ad Contrarian: The Web Is Not Medicine

mediaczar — Wed, 10 Jul 2013 10:31:38 -0000

Agree strongly. However - point of information - that list of "top brands on Facebook" is - well, a list of TV brands (so it rather begs the question.)

Here's a more balanced list: http://www.pagedatapro.com/...

It doesn't detract from your point; if anything, it offers context and support.

Re: Cameron's tweeting of the G8's luxury menu shows his blind spot

mediaczar — Tue, 18 Jun 2013 13:20:21 -0000

Don't be so sure: violet artichokes cost ~£0.90 per 100g retail. Whereas potatoes cost (depending on how many you want to buy) ~£0.95 per kilo.

What you want is brown rice. That costs around £1.00 per uncooked kilo, but will bulk up.

Re: Cameron's tweeting of the G8's luxury menu shows his blind spot

mediaczar — Tue, 18 Jun 2013 13:12:18 -0000

I do hope this link to the NHS's "Help with Prescription Costs" is of some use to you:
http://www.nhs.uk/NHSEnglan...

Re: Cameron's tweeting of the G8's luxury menu shows his blind spot

mediaczar — Tue, 18 Jun 2013 13:10:08 -0000

Seconded. Prawn cocktail. Beef, potatoes & 2 veg. Apple crumble & custard. That's pretty everyman.

Unless the OP reckons it should have been a KFC bargain bucket?

Re: Visualising Networks: Who Influences Who in the Music Industry

mediaczar — Wed, 12 Jun 2013 04:34:56 -0000

I strongly endorse your definition of influence as "someone’s ability to change another person’s behaviour."

However, I'm curious as to the way that you've applied this definition. Are you talking about "opportunity to influence" (e.g. "We suspect that Bob is often exposed to Adam's material, because Adam is on his blogroll and Twitter friendslist") or evidence of influence (e.g. "Bob cites Adam in the content of his posts")?

The latter is generally preferable (citation analysis is a pretty established field.) But I'm not entirely sure that it supports the claim that behaviour has changed (in our example, has Bob actually done something OTHER than cite Adam? Possibly this citation was simply a way of convincing his readership of something that Bob already believes? What if Bob is linking to a band's Wikipedia page? Or linking to a number of conflicting reviews of an album?)

Re: A slam dunk at FT Digital Media 2013

mediaczar — Fri, 03 May 2013 11:33:43 -0000

Interesting to note that other brands also managed to jump on the blackout story as fast (but perhaps not as well) as Oreo (see http://www.businessinsider.... for a few examples.) Some even received similar amounts of engagement and retweets -- at least, at the time.

I think Oreo was very well positioned; not only was this a strong continuation of their "Daily Twist" work, but they were prepared to invest a little media budget promoting it.

Perhaps more to the point (and this is anecdotal), they were the first to phone BuzzFeed with this story -- see http://www.buzzfeed.com/rac... (note that this post went live DURING the game, less than an hour after the tweet.)

Re: Thinking about a ‘Physics of People’

mediaczar — Mon, 15 Apr 2013 10:32:19 -0000

Excellent -- I've now bought this.

Re: Thinking about a ‘Physics of People’

mediaczar — Thu, 28 Mar 2013 12:33:06 -0000

"Psychohistory is a fictional science in Isaac Asimov's Foundation universe which combines history, sociology, etc., and mathematical statistics to make general predictions about the future behavior of very large groups of people, such as the Galactic Empire. It was first introduced in the five short stories (1942–1944) which would later be collected as the 1951 novel Foundation."

Source: http://en.wikipedia.org/wik...

Asimov was perceptive in many ways, and often had great foresight.

Re: Thinking about a ‘Physics of People’

mediaczar — Tue, 26 Mar 2013 05:52:53 -0000

All a bit reminiscent of Hari Seldon's Psychohistory...

Re: Are there any truly end-to-end social media monitoring platforms?

mediaczar — Mon, 04 Mar 2013 08:50:41 -0000

I instinctively mistrust anything that claims to be "The only xxx you'll ever need" or vendors who claim to offer end-to-end solutions.

Too often, the vendor's understanding of the domain of problems I'm trying to solve, or the specific question I'm addressing _right now_ don't match the reality.

The only end-to-end solutions I've seen in other areas of Business Intelligence tend to be very general-purpose: a database management system, a set of visualisation tools, statistical packages.

Tools in the social media intelligence sphere claim to offer so much more than your average BI package, and yet often offer so much less. The people who buy them are less clear on the briefs they need to fill, and are (only too often) happy to accept the pre-made reports offered by the tools.

It's all very depressing. But there's a ray of hope. I've seen some interesting single-purpose (or highly specialised) toolkits emerging that treat the problem differently. What if -- the social intelligence industry seems to be saying -- we do one thing well, rather than all things poorly?

Re: Are you falling foul of the Facebook 20 percent text rule on images?

mediaczar — Sat, 23 Feb 2013 02:14:22 -0000

Having checked with Facebook, it seems that their ads policy team is manually checking at present. That is, no clever software.

Re: Facebook launches Graph Search and injects search into social

mediaczar — Wed, 16 Jan 2013 07:48:12 -0000

Shameless plug for my notes: a round-up of the illuminating articles plus some pithy notes of my own...

http://blog.magicbeanlab.co...

Re: Data Science and Skyfall: Mining the opening weekend

mediaczar — Mon, 10 Dec 2012 11:55:06 -0000

This 2010 paper by HP Labs might make good reading: http://www.hpl.hp.com/resea...

In it, they go a step further and built a regression model for predicting the value of movies before their release (table 4 & fig 6).

Since then, ironically, there's been some research into whether the opening weekend is such a strong predictor of overall takings as it used to be. The argument (I believe) is that increased accessibility to / speed of publishing for peer reviews can sink a poor movie that's been well promoted...

Re: Magic Bean Lab Notes — How likely is medical debunker @bengoldacre to...

mediaczar — Sun, 02 Dec 2012 19:03:47 -0000

I really didn't look too hard; just ran through a quick list of the top pharma companies from Wikipedia (+ Cochrane) rather than trying anything smart, like term identification.

You're right about that Twitter feed thing; more to the point his comment feed is search-engine crawl-able. So it's not very robust!

But intitle:GSK OR intitle:glaxosmithkline OR intitle:glaxo site:http://www.badscience.net/ isn't a fun search.

Re: What Makes a Social Media Metric Stand Up To Scrutiny?

mediaczar — Wed, 28 Nov 2012 13:07:27 -0000

Some kind of outcome that relates to the client's core business KPIs would be nice, too.

Re: Forecasting Made Skeezy: Projecting Meth Demand using Exponential Smoothing

mediaczar — Tue, 27 Nov 2012 06:00:16 -0000

Aaah. Thank you. Just saved me £££s.

Re: Social Media and Furniture Retailers (With Added Meatballs)

mediaczar — Mon, 26 Nov 2012 11:53:48 -0000

Lovely stuff. Can you look at the differences between the audiences mentioning 'IKEA' and those mentioning 'John Lewis'?

Also, assuming that the word cloud was created using data from the same period as the rest of the analysis (September 2012), how come there's no mention of the JLP ad campaign (http://www.youtube.com/watc... that broke around the middle of that month? After all, it gathered around 25K mentions on Twitter alone in the period -- or around 500x as many as ('John Lewis' AND 'Charlie Mayfield'). Have you explicitly excluded mentions of advertising in your search terms? If so, to what purpose?

Re: Forecasting Made Skeezy: Projecting Meth Demand using Exponential Smoothing

mediaczar — Mon, 26 Nov 2012 05:44:56 -0000

Fascinating post - as ever. The intricacies of underworld analytics are gripping.

I came across a single hurdle -- which I'm sure I could work around were I cleverer -- the OS X version of Excel doesn't handle array formulae. I ended up trying this in R with satisfying results -- but results that were different from those of your protagonist.

I'd really like to follow along with the spreadsheets (it would make me a stronger Excel user for one thing, for another I'd get a better understanding of what's happening), but the differences between the OS X/Windows software make this tricky.

This isn't so much a gripe (and if it were, it wouldn't be with your post) so much as a cry for help.

Re: Federated Media kills off banner ads, shifting focus to demand-based and topic-driven advertising

mediaczar — Tue, 13 Nov 2012 06:11:19 -0000

Bad headline on good article.

FM's still saying that 40% of its revenues will come from display space -- only that they'll sell that through ad exchanges rather than invest in expensive sales staff to execute agencies' buys.

So they'll come off a CPM model, and start selling on a real-time bidding model, where each impression is negotiated on its own merits. Usually, those merits have more to do with *who* is receiving the impression (and what cookie data the advertiser has access to) and less to do with *where* the impression is being served (context.)

That may well reduce the effective overall CPM, but it should also massively reduce the cost of sales (ditching 24 sales staff, reducing marketing budgets, saying 'good bye' to agency entertainment budgets and expenses).

But it doesn't mean that they'll kill off banner ads -- or even reduce their inventory. Quite the opposite.

Re: http://magicbeanlab.tumblr.com/post/35119521859

mediaczar — Tue, 06 Nov 2012 16:03:22 -0000

This tends towards 90% if there's an associated photograph.

Re: Milestones, Tools and Tutorials

mediaczar — Sat, 13 Oct 2012 09:37:14 -0000

Thanks for this. I'm on more or less the same journey (although I've really not begun to look properly at anything Javascript-like -- I sense this may be a mistake.)

Re: http://magicbeanlab.tumblr.com/post/32658651230

mediaczar — Thu, 04 Oct 2012 16:50:52 -0000

I think that their data's useful. And I understand there's pressure on them to publish "sensationalist" stuff to attract press attention around the release. But I think they might be editorialising too much.