Jul 18, 2016

The stone age of bots

All of us are excited about bots, but we don't have our "PokemonGo" moment yet, the one bot everyone gets batshit crazy about. So it is hard to pitch bot ideas to management when, if you were honest, in most business case columns you would have to write "we have no clue". You can rather win budgets with "gathering insights, learnings, experiments" and "it won't do harm to our core business" currently. Evidence? Try out these: "8 bots you should add to your Facebook messenger app". And then tell me which one is great:

http://linkis.com/www.makeuseof.com/ta/2v3sC 

None is. The long term vision may be that people will text businesses rather than call them. And that not a person, but a computer will answer them. I am excited about the possibilities, too. But until the computers texting back aren't smarter than people - "if you want this, press one" - until then, letting people talk with people will be a major customer service advantage. Check for example the experiences of Indian heavyweight "Helpchat". I mean, it's even in their name, but they clearly found out where a chat helps - and where not.

https://inc42.com/resources/moving-away-chat-hard-earned-lessons/

The tragedy of PokemonGo

Interesting article on The Atlantic that puts PokemonGo into a gaming history perspective. I personally believe that, more than "augmented reality" and "location based gaming", the Pokemon franchise is the reason for its success. People who are 32 today have had great afternoons and evenings playing Pokemon on their 90s consoles. Kids know the franchise from their Nintendos. it is just the perfect brand and set of characters to move Nintendo away from the consoles and onto the mobile phones, although Niantec probably did the hard part of the moving: But without the Pokemons, it wouldn't work that well - it would just be another interesting experiment and not such a mass phenomenon.

http://www.theatlantic.com/technology/archive/2016/07/the-tragedy-of-pokemon-go/490793/

But maybe Pokemon will be the ice breaker for similar apps or augmented reality in a business context as already now I am being asked what to learn from PokemonGo's success, but for now, after having the game for a week, I guess the answers are a lot simpler than one would expect:

1. The power of an established "love"-brand

2. The power of the consumer - even if you plan to launch in different countries subsequently, they will simply find the APK from other countries and start playing

3. Don't ignore the history of games that have paved the way, this success does not come out of nowhere (esp. in game concept/tech side), see link

4. Don't underestimate, even in 2016, the tech & server power needed for realtime, location based games on a global scale

5. Don't think that without any foundations (Niantec's Ingress database & experiences, Nintendo's Pokemon brand) you could build a similar app for your bank or chain of supermarkets and could put "if we only reach 4% of PokemonGo players" in your business plan excel sheet, you would succeed




Jul 4, 2016

Danish agency uses AI to make advertising (media) decisions

There has to be something about Denmark and Mathematics. If you like football (not "handegg" like they play in the US), the name "Midtjylland" probably rings a bell. Yes, that Danish club that made it to the group stage of the UEFA Champions League. Its owner, Matthew Benham, made his fortune with mathematical models predicting the results of football matches. He started to manage the club on statistical models and was vastly successful, winning the league.

Now we hear of Blackwood Seven, a media agency founded in 2013 in Copenhagen. It is not really a media agency, though: Thy consider themselves a software and analytics platform, charging their clients a software fee, no kickbacks, No media commissions, no gut feelings: You feed the algorithm with your KPI's and budget, and the thingy uses predictive modeling to try and determine the best media mix incl. TV, OOH, and of course digital. This approach won them the huge Volkswagen account in Germany, and it threatens the whole way media agencies make their business. Not only the kickbacks and the special rates which are often not disclosed to clients - meaning companies like Blackwood Seven make the whole market more transparent - but also when it comes to media planning decisions. When computers beat the best Chess (long time ago) and Go (recently) players in the world, why shouldn't they, fed with Big Data, beat your media buyer, who most probably isn't the best in the world?

More about Midtjylland:
"How data, not people, call the shots in Denmark"
https://decorrespondent.nl/2607/How-data-not-people-call-the-shots-in-Denmark/230219386155-d2948861

https://www.theguardian.com/football/2015/jul/27/how-fc-midtjylland-analytical-route-champions-league-brentford-matthew-benham

German: http://www.11freunde.de/artikel/midtjyllands-revolution

More about Blackwood Seven:
https://blackwoodseven.com/

http://www.businessinsider.de/blackwood-seven-launches-in-the-us-2016-6?r=US&IR=T

The messenger platform gets an update - and a blog

Messenger bots are for now not as exciting as we have imagined them. It is so important to remember how new the whole thing is, although we have seen chatbots around for more than one and a half decades now. But given some time, we will have a few "blockbuster" bots. In the meantime, Facebook continues to develop the platform and, because there seems to be a lot of need for communication, a blog to let businesses and developers follow closely what they could do on Messenger.

Update/new features:
https://www.facebook.com/notes/david-marcus/messenger-platform-gets-an-update/10155014173359148

Messenger blog:
http://messengerblog.com/

The Washington Post is a tech company now

Remember when Jeff Bezos came and we all wondered what he would do with the Washington Post? A few years later, the company has a CMS from which it not only powers all its digital products, it licenses it to other publishers and someday wants to make 100 million USD from this business alone.
"I want the NYT to call me and say 'Holy shit, I want that', says the Post's head of ad product and technology, who also produces products that will solve ad problems for other publishers, too. Interesting read, especially when yuo notice how little this whole article talks about journalism:

http://nymag.com/daily/intelligencer/2016/06/washington-post-jeff-bezos-donald-trump.html#




[German] 30 Stunden Woche - scheint zu funktionieren

6 Stunden Arbeit am Tag. Das klingt für viele wie "Ferien" oder "Wochenende", ist aber offenbar ein Modell, das man sich genauer anschauen sollte - denn die Produktivität scheint in vielen geeigneten Arbeitsbereichen die des 8-Stunden-Arbeitstages zu übersteigen. Aus meiner Sicht sehr gut vorstellbar, dass das funktioniert. Häufig sehe ich - außer Manager, die den ganzen Tag durchgeplant sind und in Meetingräumen sitzen - viele Menschen, wenn ich denn mal Unternehmen besuche, die zwar 8 Stunden im Unternehmen sind, aber de facto vielleicht fünf bis sechs wirklich arbeiten, und das oft mehr schlecht als recht - denn "das Unternehmen" ist eben ein Zeitfresser, macht Hobbies fast unmöglich, lässt einen nur am Wochenende mal eine Boutique von innen sehen usw. Wissenschaftler argumentieren wohl schon länger so und sagen, dass sich niemand (oder nur sehr wenige Menschen) 8 Stunden lang wirklich konzentrieren und auf hohem Niveau arbeiten kann. Konzentrierte sechs Stunden plus einer halbstündigen Pause in der Mitte - kann ich mir als ein spannendes Modell vorstellen. Die Schweden auch:

http://ze.tt/goeteborg-testet-den-6-stunden-arbeitstag-mit-ueberraschenden-ergebnissen/

You need 17,000 units (and a few streams) to be No. 1 on US Music Charts

Rihanna's album "Anti" is No.1 on the US "Billboard 200". Overall, it says there were 54,000 units moved last week, which is low already for probably the biggest music market in the world. But since last year, they add up streams and units (I guess physical and download) = 1,500 plays of an album count as "one unit sold" apparently. In Rihanna's specific case, the breakdown shows 17,000 units sold, the rest comes from streaming. Forbes says it's the lowest selling No. 1 ever.

http://www.forbes.com/sites/hughmcintyre/2016/03/20/this-weeks-number-one-album-is-the-lowest-selling-in-history/