Aggregate cost-per-click (11)% (1)%
Cost-per-click on Google
websites (16)% (2)%
Cost-per-click on
Google Network Members' websites (4)% 1%
That is the search advertising margin eroding. Google has been compensating for that in a number of ways, from putting more ads on their sites (more "inventory") to re-writing rules ("no mobile for third parties") to changing splits with non-Google sites. The underlying cancer though is that Advertisers are more and more likely to see Microsoft Bing, Facebook, and Yahoo! ads as "equivalent" to Google's ads. And they split their spend, and they have been less and less impressed with the "results" of their advertising, so they pay less. All symptoms of a maturing market and one that has been going on for years. And as it continues it puts more and more pressure on Google to make up its margins in other ways. My guess is that 2016 is the first year they are going to have to cut staff across the board to keep their margins up.
So you're confounding average CPC with search advertising margin. Great. And then making a bunch of statements about the cause for a decrease in CPC with no factual basis. Here's another plausible narrative:
Mix of traffic is shifting as traffic builds in new platforms (e.g. mobile, emerging markets, YouTube), knocking down average CPC. Then we're getting a Simpson's paradox (https://en.wikipedia.org/wiki/Simpson%27s_paradox) effect causing average CPC to decrease.
Actually not so much, there is a fun class they used to give new employees at Google (called Nooglers) called "life of a dollar" (well that was what it was called when I joined) which talked about how Google made money with search advertising. And one of the key things about that class was how the cost per click, or the amount of money Google got when someone clicked a link, was fundamentally tied to how all of the other processes made (or didn't make) money. After I left Google four years later, I actually joined another search startup called Blekko where not only did I get to see the "high level" view of where Google made its money (and how much it would share with us if we used their ads) I also got to see the contracts involved and the clauses where Google wanted to protect its margins. We also partnered with one of Google's largest advertising redistributors, Infospace.
When I left in 2010 I started also following Google's financial reports. Year after year, following that indicator you could see pressure on Google's search ad business, and you could watch Microsoft's Bing ads get to be more and more profitable. Did you know that if in 2010 Microsoft could get the same CPC as Google did for their ads on Bing that Bing would have out performed all other parts of Microsoft in terms of revenue and profit? As a search vendor in the middle I could see what Microsoft shared with us and what Google shared with us, and we could back compute the actual CPC, and the ratio of the CPC. Whats more we could watch Microsoft's going up while Google's was (and is) continuing to go down.
I was also there as their CPC plummeted and they told Infospace (and through them us) that we could no longer show Google ads on Mobile search. No only Google could from now on. And that wasn't just us, that was everyone. And we've watched as more and more of their search page became "sponsored" in one form or another, and how people with Google ads on their web sites were getting less and less money, even though their traffic was the same or better.
I've been watching the Google money machine slowly deflate for the last 5 years (coming up on 6) and called almost every move they've made prior to them making it. All based on understanding that rounded to the nearest billion dollars, nothing in Google makes any money except search advertising. Even Youtube can barely break even these days and we've seen how the 99% of Youtube "stars" actually have real day jobs rather than be able to produce content for that service. We've seen Google working hard to "yank back" the whole Android is open source thing, in the name of "user experience" to be sure, but attacking Cyanogen? Threatening Samsung? Forcing HTC to throw itself under the Microsoft bus? And paying for more and more traffic, $4 billion a year? how much further can they push that life vest?
The good news here is that predictions are like freebies, you win some you lose some. No alien contact this year, sorry Inquirer! But I've been more right than wrong about predicting Google's road and I'm confident in this one too. Google is stuck between a need to please the street and the cost of engineers that produce no revenue. They have way more of the latter than they need, and from what I can tell they have pretty much been aggressive as they can at getting rid of people "gradually" who were in the bottom 10%. And their CPCs continue to fall double digit percentages year over year, and they don't have any business that is growing at that rate. Throw in an EU anti-monopoly monkey wrench, or a Facebook throwing their search over to Bing, or the EU getting their tax policy together and eliminating the double irish? Suddenly the need to shed costs fast will overwhelm their ability to do so quietly.
I kind of thought they would end up doing that this year but they dodged a bit by turning into Alphabet which gave them some new accounting tricks.<...
TL;DR: you worked at Google where they said average CPC was important and you worked on a search related startup that sold relabeled Google ads (presumably through the AdSense for Search program?) where Google paid some contractually fixed fraction of revenue on searches (revenue on these searches is not necessarily correlated with revenue on main Google search, correct?). At this startup, they wouldn't let you resell these ads for mobile searches because it would hurt margins (seems like shaky reasoning to me, who doesn't like more money / more profit?).
Then a smattering of recent news stories with tangential relationships to your main thesis about margins (YouTube (when have the margins been broken out on this?), Android (not sure how this fits into your narrative about the decline of Google's growth)) mixed in with some real medium term competitive threats (EU, Facebook/Bing (didn't FB used to redirect their search traffic to Bing? not sure that made much of a difference)).
Here's some alternative narratives:
a) There's far more to profitability on online advertising than average CPC, like query mix as I mentioned before. Additionally, advertising on Google owned & operated properties have been continually far more profitable and growing at a much faster pace than relabeled search advertising or AdSense on third party websites. Google has refocused their engineering / sales efforts on advertising products on their owned & operated properties and de-emphasized properties like Blekko.
b) Google's PE ratio has gone from <20x in 2010 to >30x today. The market believes Google's growth prospects are brighter than ever.
To summarize, I also keep a careful eye on Google / other tech company's financials and how they relate to both my personal work and larger trends in the industry as a whole. I simply disagree with your theses that a) search engine margins are decreasing and b) that the search engine business needs to grow at a faster rate for Google to continue to grow sustainably due to a) the continued power of direct response search advertising and b) other growth areas like video, brand, mobile, and new markets.
I was with you until nr. 7. Definitely not going to happen. If some genius makes it happen, expect it to be an epic failure.
8 and 10 maybe, 9 I don't know...
I would be interested in your reasoning for why it would fail. All major arms makers have a demonstration biometric activated weapon, and there is a lot of pressure in the US on private gun ownership. I see these forces coming together to force biometrics as a requirement for owning a new gun.
(My reply is specific to this happening in 2016. There will one day be Judge Dredd-like firearms. I don't see that being widespread in my lifetime, though.)
If you work in a company developing a hardware product you will learn, amongst other things, this:
1) The demonstration item is very likely hacked together at the last moment, and would 100% fail in the market if it was put into production at the time of the demo. (Even if your company does everything right, you are frequently only a part of the total product).
2) You never use the latest, shiniest, unproven technology. You use the simplest, production-friendly technology that will perform the job. Otherwise you are building a product higher than necessary failure rates.
(This is assuming that you aren't working for Google, Apple, Elon Musk or someone else who can put 100+ engineers on a sub-project and/or invent new meterials etc).
My assumptions are thus:
1) The manufacturers probably did the demonstrations to a) not be outdone by the competition and b) gauge demand.
2) No arms manufacturer has a robust, proven design that is affordable for civilian small arms. Going out on a limb I would also guess that none of them have either an experienced embedded engineering team or an intimate, effective working relationship with an external team. This is a significant expense (as opposed to a mockup for demonstration purposes), from which (3) follows:
3) The only way that a majority of customers will pay extra for this (anti-)feature is if legislation forces them to. California (and a few other states?) might try this, but given the NRA's lobbying power (and other factors) it won't be federal. If California tries it, it will be only for new firearms and perhaps even only for a certain group of sales/type of firearms, as a test. I don't see many such firearms being sold in the next 20 (at least) years.
Now, everything up to here has been some nice mental masturbation, rendered moot by the reality of a biometric-restricted firearm. These problems are non-exhaustive and can be solved, but it won't be cheap, easy or quick to market.
1) When the cheap "fingerprint" sensor with barely functioning, insecure firmware in your phone fails, no problem. If that happens on a firearm, big problem.
How can it fail? Battery runs out. Sensor gets smashed. Blood on your fingers. Finger cut/crushed AND blood on it. / Gelatine fingerprint. Real finger severed by criminals. Attacker owns a screwdriver.
2) RFID.
How can it fail? Battery runs out. Backup battery runs out. Tag gets smashed. RFID frequency gets jammed. / Tag gets recorded by someone walking past you. Tag gets recorded by someone with a directional antenna. When strong encryption is finally added, the implementation is flawed and gets cracked. One of numerous other RFID issues gets exploited (see Defcon, Chaos Communication Congress et al.). Attacker owns a screwdriver.
3) Dynamic Grip Recognition
How can it fail? Owner weakened by attack, fatigue or medical condition. Dominant hand damaged.
4) Conversion kits using any of the above
Since the aren't integral to the weapon, disabling or removing them will likely be very easy.
5) Source code of the firearm
You want to put a binary blob on my firearm? Firstly, no. Secondly I envision some interesting "unintended acceleration"-like lawsuits. (That reportedly cost Toyota about $1000 per line of source code)
Addressing all of these issues will take a long time and will be really expensive. And I'll bet the market will then end up with a DRM'd firearm that the owner won't really own. Maybe even connected to the internet, with a GPS and geographic fencing. There is a final addition which I would actually support: a screen and loudspeaker. Because then you can force the owner to view educational material on firearm safety, anger management, conflict resolution, the results of employing violence, responsible drug use and where they can find counselling w.r....
Not sure what triggered the filters on your response @random but I appreciate you taking the time to write it out. I can see where your scepticism comes from and agree with you that its a stretch. We'll have to see if the gun manufacturers can pull it off or not.
A couple of comments on your comment, at the 2012 SHOT Show there were people showing biometric trigger locks and gun safes, last year the iP1[1] (RFID) was there. At that time both Sig and Glock indicated they would have something similar this year. We'll know in a bit more than 20 days :-).
If I understood the essence of your argument, it was that this is a hard problem that gun manufacturers are not investing in solving, and under investing will only yield ineffective solutions.
I can see the point, I'll wait to see what people are showing next month before I cede my prediction though :-)
I was wondering why 8 isn't more wide-spread. We tend to see all these fun applications in tech news but the unfortunately obvious route for UAVs is for war/terror purposes (beyond the predator drones). On a smaller scale, these will be used as a replacement for suicide bombers and even targeted kills.
Because it still isn't cost effective. A drone that can be guided, either autonomously or by a human, will cost at least, what, $80 wholesale? That will almost equal the price of the ordinance. Causing a lot of casualties at a low cost is one of the major appeals of field artillery and explosives.
Now, maybe it would make more sense for the assassination of carefully guarded targets. But it will probably be a few more years before that technology is more reliable than strapping a vest containing c4 on someone.
That works when you are in 'gun range' of their cover and can shoot at them. But if you're pinned down? What about the enemies rear area where they are staging? How quickly can you set up and tear down a mortar emplacement?
Giving the soldier on the ground the ability to send a round exactly where they want to, behind cover, even inside a building. And I think you've created a very strong force multiplier. Given the the current state of the art in having locally operated visual recognition on the drone itself (http://www.csail.mit.edu/drone_flies_through_forest_at_30_mp...) to either avoid obstacles or target specific objects makes the weapon even more effective.
In my opinion when these things make their debut they will be as disruptive as the machine gun.
I see tiny surveillance drones on every helmet, with HUD controls. Look into every room before entering; see what's coming up around every bend. And land back on the helmet to recharge! Like the SAW, need one on every team.
Random predictions from the top of my head:
1. Funding for start ups is lower than 2015
2. Genomics and anything related to autonomous vehices (including drones) will be the hottest sectors
3. Google's revenue from non ad based products will double
1) React.js starts to lose popularity due to it's ultra complex tooling ecosystem. People want to feel like what they learn will be still be useful at their next job. The React ecosystem doesn't provide that; the React ecosystem tends to burn people out.
2) Smaller frameworks like Vue or Riot takes the spotlight.
3) Angular 2 is a hit thanks to its "batteries included" design, which will appeal to React burnouts.
4) A new JavaScript rendering-based framework will come out and become a hit. Something like Turbolinks or Glimmer, except it doesn't break jQuery or sandboxes you to an ecosystem.
Alternative prediction: the React.js ecosystem starts to standardize and stabilize, probably around Redux and WebPack. React developers' productivity and satisfaction improves.
> 5) JavaScript loses popularity, MEAN stack loses popularity but alternatives aren't appearing, just more criticism of the javascript frameworks
I think, there will be alternatives. WebAssembly will let developers write client-side frameworks in C/C++, C#, Go, etc. Very likely, JS will start losing its popularity as the selling point "works both on client and server" will work for any language. However, I don't believe we can say goodbye to JS. At least, not in 2016.
1. Battery powered automotive devices, both small(hoverboards, drones) and large(cars), continue picking up steam; articles circulate about the potential danger and disruption of the new devices
2. Multiple new operating system projects announced, targeted for unikernel VM deployment
3. Emerging AI techniques commercialized for creative applications, e.g. a
new wave of selfie apps, professional art or music tools
4. Distributed apps using blockchain protocols start making small ripples
5. Software and devices that successfully bridge the mobile/desktop gap are demonstrated
6. VR/AR devices ship, but demand remains modest and mostly in professional niches
I would have said something about the economy and finance too, but this is a tech predictions thread. So I'll go for a tech-economic one:
7. Trends turn against one or more of the current leading social networks, as a bold newcomer finds an opening
8. Bubble mania in the Valley peaks and shifts towards panic as key macro indicators start sagging
Regarding #8, has a situation ever occurred where a large majority of people assumed there was a bubble, reacted accordingly, and created negative outcomes, despite there not being a bubble?
And I hope (7) Twitter's death will be quick, and not painfully slow.
This is something I have been hoping for or at least curious for a long time - we're able to specify script language but there is really only one supported language
Please not, Lua would be easy to implement but if we can only pick one language then it's too similar to JS. We need more diversity, static types+AOT+noGC would be a good start.
Also if we do get Swift it will be for the same reason we got <canvas>. Apple wanted to make dashboard, so they put an extension in webkit, people started using it on the open web and...
Lua would certainly be a good choice. A really good language syntax, the language is small, and LuaJIT2 is very very fast.
Though, it's better to stay with one language, JavaScript - it's a good language, and everyone should learn it first (I would argue most simply to don't understand Javascript and use only a fraction of it, in an non-optimal way). Fragmentation is bad for the web. I would already consider CoffeeScript/TypeScript/Dart/etc that compile to JS bad because its fragmenting the developer base and it makes it harder to reuse code. Have you ever visited a GitHub page only to find out that a lovely open source project is coded in some obscure language that compiles to Javascript.
1) Deep Learning-based techniques outperform hand engineered natural language processing stacks in every measurable way.
2) There will be considerable progress made on the Winograd Schema challenge.
3) The (unconstrained) Turing test won't be "officially" passed until October 2017 though (it will actually be passed in July or August, but the news won't leak until October)[1].
4) Some people will continue to insist that Deep Learning is nothing different to what was being done in the 1990s. At some point someone will get frustrated enough with this to blog everything that is different now.
What do you mean by hand engineered? Is a part-of-speech tagger trained on a manually annotated corpus hand engineered? If not, then which "stacks" are hand-engineered and still popular today?
Yes this is true. Perhaps I should better express it as a hand engineered feature pipeline, ie the way people these days will POS tag, then lemmatize, then do something with WordNet, then maybe some Semantic Frame thing and finally have an application they can benchmark. Writing that pipeline is manual, and it has hand engineered features in it too, though some may be trained.
1) Ad blocker usage doubles on the desktop in Western Europe and the US by the end of the year.
2) Websites locking out users of ad blockers becomes routine, rather than exceptional.
3) There will be at least one successful legislative attempt to outlaw ad blocking, in Europe.
4) Long shot: Apple ships a minor (+0.1) update to iOS with ad blocking enabled by default. (Blackberry might do the same, but it would not be as consequential as Apple doing it.)
3) Why Europe? The US and Australia have more powerful copyright and publishing lobbies and far worse laws as a result. Adblock Plus has already successfully defended itself from publishers in court in Germany.
I think the pro-user lobby (EFF, Wikimedia Foundation, Internet Archive and others) is strong and organised enough in the US that I don't think such a law would be politically possible today. No individual European country has an equally effective equivalent that I know of.
I hadn't considered Australia, but you are right, that would be a good candidate as well.
Sorry, I hadn't seen your comment til now. Well, it'd work just like Meetup.com, except it'd probably be an app, or at least have a mobile component. Everyone would have a profile with their interests and when two people in the same geographic area with the same interest(s) sign up, the system would alert them both so they could set up a one-on-one meetup. There'd also be an optional GPS function where if you're free (w/ nothing to do) and wouldn't mind having a good conversation with someone regarding your interests, the app could ping you when two similar users are near each other.
1) FinTech will start to shine through, the first consumer banks built using modern tech will open to customers in Europe (London FinTech scene is strong, Zurich has ex-Googlers and strong finance, or Frankfurt but they are currently trailing - my bet is on London). These banks will experience very strong growth, the question really is: Will they go with it or go for acquisition
2) The "Family Plan" will emerge as a new sales market in most of the established consumer products, with Dropbox, Google, and others all building strong offerings for managing the product use and sharing of a group of users. This is ground-work for centralising both "family" and "home", and lays the foundation for the command and control of IoT over the next few years - it is how the big established players stay in the game.
Neither of those predictions is quite there, the leading new bank is Mondo https://getmondo.co.uk/ but it's in beta, is currently iPhone only which limits adoption in the UK, and the benefits of the new tech hasn't yet been fully realised.
And I'm not yet seeing the "family plan" head this way but I'd be surprised if the penny doesn't drop somewhere and this be the path taken. Control of the home is control of the family, and vice versa... the family want tools for this better than the ones they have today.
An entirely different thing I'd like to see exist but do not know of anyone working on at all:
3) A dating app that acknowledges the hook-up culture that seems to be growing in the millenial generation of users, and instead of hiding that under the carpet uses it's data to encourage responsible tracking of STDs and other risks.
That one is inspired by news this morning on Gonorrhoea becoming immune to antibiotics http://www.bbc.co.uk/news/health-35153794 . The app should function similar to whatever the core function of Tinder is, but allow tracking of whom you've done what with for the sole purpose of allowing notifications of STDs to be quickly disseminated to people who may be exposed.
It may be infeasible, I don't know what the willingness of those who do hook-ups is to track risk-related activity and test results is. The core idea is "disease tracking in social networking and dating apps".
On the banking thing. In NL we have Bunq which is a virtual bank with banking license. It combines social with banking. The idea is nice but I'm not sure if the world is ready yet. https://www.bunq.com/en/
I like that they have one of the things I anticipate will emerge "Instantly switch your debit card between accounts".
I hope to see features that allow the management of accounts by splitting an account into virtual buckets, allowing transactions from frequent providers to be tagged/labelled in some way.
Or even to map retailers/companies such that "transactions from them are bills and are always associated with this account", even though you only carry the one debit card it will act intelligently as simply a payment card and you would have control over where the money actually comes from.
Can you elaborate on what is the value proposition/differentiator for such a "new tech" consumer bank? I'm not following this sector closely and are curious about what's brewing.
The traditional banks have very slow systems, and those prevent a lot of new scenarios.
For example:
* When travelling, you could use a mobile app to receive instant notification of every payment made, which helps you manage your fraud risk when abroad (temporarily block the card if you see something suspect, whilst allowing you to whitelist for specific transactions so your card actually still works).
* You could have control over the "routing" of your finances... designate payments as bills and have them come out of a different account than the one you use for disposable income.
* You could split payments, go for a meal and one person pays and then agree to split and the payments are auto-managed amongst those who were there (similar to Ubers split fare)
* You could have better budgeting and management of finances, better reporting, as the banks could run the equivalent of "Google ID" (Google have this for phone number recognition) for payments so that transactions are automatically tagged and managed.
* You could have dynamic control of spending limits, set alerts and controls that follow your normal pattern to really increase fraud prevention measures, whilst your app would allow you instant control when you need to make purchases outside of that pattern.
Really it's about putting consumers in control of their money in all ways.
Banks are horribly slow, and the slowness removes all control.
Edit: I should add, there is a driving force behind this in Europe: Contactless card payments.
In Europe we've had contactless "touch" payments for several years, they are ubiquitous and can be used for any payment below £30. They are now so common that most people I know carry no cash at all, they simply touch their debit cards and the transaction is done in seconds.
This has led to very noisy bank accounts, where a bank account used to have fewer than 50 transactions per month: bills + cash withdrawals, the bank accounts now have almost every transaction: a stick of chewing gum, a sandwich from Pret, a coffee in the morning.
The greater the noise, the more speed and control are real benefits to the consumer.
On the other hand, traditional banks can easily copy all these features if they become important to customers. I'm not sure they can be long-term differentiators - unless software patents, which fortunately here in Europe are much less strong than in the US - and I don't see any network effect to give the first mover a strong advantage.
1. Windows 10 usage ratio explodes. Linux distributions might see a small increase on the desktop market (up to something like 2%).
2. Major browsers will start experimenting with warning the users that they are visiting a site that does not support HTTPS (although I think that they won't be adding that to stable versions until 2017).
3. Google+ is going away for good.
4. European Union makes the progress in uniting as a single market to fight against geo-blocking.
5. Self-driving cars are still not ready to be commercially available.
1. Privacy becomes more important to social networks
2. Social networks withdraw more APIs to protect not only privacy but their own business interests
3. Privacy becomes more annoying to lawmakers
4. Lawmakers in the US vote in more surveillance powers to force social networks to comply with their information demands
5. Lawmakers in the EU create new laws to constrain their citizens' data to EU countries
6. Distributed social networks pick up traction in both the EU and US in response to #3 and #4. One will be promoted by a major player, probably Google as a play against Facebook.
One year is not really that long - but here is what I am seeing in the foreseeable future:
1. Development for proprietary mobile platforms (such as iOS/ObjectiveC/Swift and Android/Java) will drop in popularity.
2. Continuing increase in client-side JavaScript/HTML5 popularity. The dominant platform will be ES6/React not TypeScript/Angular2.
3. The "App Store" will start to loose its relevancy.
4. Mobile (vs. web) usage continues to grow.
And for business:
5. Social networks coming out of Asia (China, Japan, Korea, SEA - such as WeChat or Line) becoming globally popular.
6. Crash/collapse/accounting fraud scandal in a major internet company. (No idea for specifics but look at the ride subsidies of uber for an example of something that can go wrong).
7. Xiaomi starts selling in USA and Europe with considerable success.
> 2. Continuing increase in client-side JavaScript/HTML5 popularity. The dominant platform will be ES6/React not TypeScript/Angular2.
My prediction is a better alternative to JS for client-side development. WebAssembly, I'm pointing at you. And as a result, a lot of new client-side frameworks.
1. iPhone7 will finish the android dominance finally. Apple Pay become prevalent.
2. Google will re-license android so only Google can make android phone. Chinese and Samsung will either stuck with the last android OS or pushing into a fork.
3. Driver assistance AI will become standard in 2017 model car, we will see car industry the most active AI research and acquisition market.
4. AMD and nVidia competition heat up and will drive deep learning into deeper and ever increasing number layers, and hit a breaking point when the next Google emerges to replace search and Wikipedia, at finger tip.
5. Microsoft Hololens will be a hit, it actually revives Microsoft from doom. But we will find Amazon enter into VR market.
6. Amazon drone delivery will start in a limit number of area, suburban mostly.
7. Elon Musk will bankrupt as the oil price continue in downward trend, and solar panel adoption slowdown as punished by power company.
8. We will see the first Bitcoin/blockchain based crowdfunding, may even be equity based. Meanwhile Bitcoin price will double as bankers busy push their own blockchain based banking infrastructure, while ignorant to Apple Pay is eating them alive.
9. Facebook enter China market, and increases its size of user by another quarter, while Facebook breakthrough in online translation for the first time makes choice language irrelevant.
10. Trump wins presidential election, WWIII becomes the predominant theme of the computer games released in the years followed.
112 comments
[ 4.5 ms ] story [ 84.3 ms ] thread2) The first re-use of a rocket booster to launch a payload into orbit.
3) A commercially exploitable use for Graphene is found.
4) Some genetic condition is completely cured in mice using CRISPr techniques.
5) VR/AR actually ships in underwhelming quantities.
6) Power companies sue to block people from installing whole house batteries.
7) Biometric firearms see widespread adoption.
8) Drone "mortar" shells (single use drone carrying a shrapnel grenade) see use in the battle fields of the middle east.
9) Google has its first wide spread layoff not associated with an acquisition.
10) Nintendo ships a fun to use game console.
Margins went up YoY in Q3 2015: https://investor.google.com/earnings/2015/Q3_google_earnings...
Mix of traffic is shifting as traffic builds in new platforms (e.g. mobile, emerging markets, YouTube), knocking down average CPC. Then we're getting a Simpson's paradox (https://en.wikipedia.org/wiki/Simpson%27s_paradox) effect causing average CPC to decrease.
When I left in 2010 I started also following Google's financial reports. Year after year, following that indicator you could see pressure on Google's search ad business, and you could watch Microsoft's Bing ads get to be more and more profitable. Did you know that if in 2010 Microsoft could get the same CPC as Google did for their ads on Bing that Bing would have out performed all other parts of Microsoft in terms of revenue and profit? As a search vendor in the middle I could see what Microsoft shared with us and what Google shared with us, and we could back compute the actual CPC, and the ratio of the CPC. Whats more we could watch Microsoft's going up while Google's was (and is) continuing to go down.
I was also there as their CPC plummeted and they told Infospace (and through them us) that we could no longer show Google ads on Mobile search. No only Google could from now on. And that wasn't just us, that was everyone. And we've watched as more and more of their search page became "sponsored" in one form or another, and how people with Google ads on their web sites were getting less and less money, even though their traffic was the same or better.
I've been watching the Google money machine slowly deflate for the last 5 years (coming up on 6) and called almost every move they've made prior to them making it. All based on understanding that rounded to the nearest billion dollars, nothing in Google makes any money except search advertising. Even Youtube can barely break even these days and we've seen how the 99% of Youtube "stars" actually have real day jobs rather than be able to produce content for that service. We've seen Google working hard to "yank back" the whole Android is open source thing, in the name of "user experience" to be sure, but attacking Cyanogen? Threatening Samsung? Forcing HTC to throw itself under the Microsoft bus? And paying for more and more traffic, $4 billion a year? how much further can they push that life vest?
The good news here is that predictions are like freebies, you win some you lose some. No alien contact this year, sorry Inquirer! But I've been more right than wrong about predicting Google's road and I'm confident in this one too. Google is stuck between a need to please the street and the cost of engineers that produce no revenue. They have way more of the latter than they need, and from what I can tell they have pretty much been aggressive as they can at getting rid of people "gradually" who were in the bottom 10%. And their CPCs continue to fall double digit percentages year over year, and they don't have any business that is growing at that rate. Throw in an EU anti-monopoly monkey wrench, or a Facebook throwing their search over to Bing, or the EU getting their tax policy together and eliminating the double irish? Suddenly the need to shed costs fast will overwhelm their ability to do so quietly.
I kind of thought they would end up doing that this year but they dodged a bit by turning into Alphabet which gave them some new accounting tricks.<...
TL;DR: you worked at Google where they said average CPC was important and you worked on a search related startup that sold relabeled Google ads (presumably through the AdSense for Search program?) where Google paid some contractually fixed fraction of revenue on searches (revenue on these searches is not necessarily correlated with revenue on main Google search, correct?). At this startup, they wouldn't let you resell these ads for mobile searches because it would hurt margins (seems like shaky reasoning to me, who doesn't like more money / more profit?).
Then a smattering of recent news stories with tangential relationships to your main thesis about margins (YouTube (when have the margins been broken out on this?), Android (not sure how this fits into your narrative about the decline of Google's growth)) mixed in with some real medium term competitive threats (EU, Facebook/Bing (didn't FB used to redirect their search traffic to Bing? not sure that made much of a difference)).
Here's some alternative narratives:
a) There's far more to profitability on online advertising than average CPC, like query mix as I mentioned before. Additionally, advertising on Google owned & operated properties have been continually far more profitable and growing at a much faster pace than relabeled search advertising or AdSense on third party websites. Google has refocused their engineering / sales efforts on advertising products on their owned & operated properties and de-emphasized properties like Blekko.
b) Google's PE ratio has gone from <20x in 2010 to >30x today. The market believes Google's growth prospects are brighter than ever.
To summarize, I also keep a careful eye on Google / other tech company's financials and how they relate to both my personal work and larger trends in the industry as a whole. I simply disagree with your theses that a) search engine margins are decreasing and b) that the search engine business needs to grow at a faster rate for Google to continue to grow sustainably due to a) the continued power of direct response search advertising and b) other growth areas like video, brand, mobile, and new markets.
I'd rather they didn't even try. Their iOS foray felt like a step in the right direction to me.
If you work in a company developing a hardware product you will learn, amongst other things, this:
1) The demonstration item is very likely hacked together at the last moment, and would 100% fail in the market if it was put into production at the time of the demo. (Even if your company does everything right, you are frequently only a part of the total product).
2) You never use the latest, shiniest, unproven technology. You use the simplest, production-friendly technology that will perform the job. Otherwise you are building a product higher than necessary failure rates.
(This is assuming that you aren't working for Google, Apple, Elon Musk or someone else who can put 100+ engineers on a sub-project and/or invent new meterials etc).
My assumptions are thus:
1) The manufacturers probably did the demonstrations to a) not be outdone by the competition and b) gauge demand.
2) No arms manufacturer has a robust, proven design that is affordable for civilian small arms. Going out on a limb I would also guess that none of them have either an experienced embedded engineering team or an intimate, effective working relationship with an external team. This is a significant expense (as opposed to a mockup for demonstration purposes), from which (3) follows:
3) The only way that a majority of customers will pay extra for this (anti-)feature is if legislation forces them to. California (and a few other states?) might try this, but given the NRA's lobbying power (and other factors) it won't be federal. If California tries it, it will be only for new firearms and perhaps even only for a certain group of sales/type of firearms, as a test. I don't see many such firearms being sold in the next 20 (at least) years.
Now, everything up to here has been some nice mental masturbation, rendered moot by the reality of a biometric-restricted firearm. These problems are non-exhaustive and can be solved, but it won't be cheap, easy or quick to market.
1) When the cheap "fingerprint" sensor with barely functioning, insecure firmware in your phone fails, no problem. If that happens on a firearm, big problem.
How can it fail? Battery runs out. Sensor gets smashed. Blood on your fingers. Finger cut/crushed AND blood on it. / Gelatine fingerprint. Real finger severed by criminals. Attacker owns a screwdriver.
2) RFID.
How can it fail? Battery runs out. Backup battery runs out. Tag gets smashed. RFID frequency gets jammed. / Tag gets recorded by someone walking past you. Tag gets recorded by someone with a directional antenna. When strong encryption is finally added, the implementation is flawed and gets cracked. One of numerous other RFID issues gets exploited (see Defcon, Chaos Communication Congress et al.). Attacker owns a screwdriver.
3) Dynamic Grip Recognition
How can it fail? Owner weakened by attack, fatigue or medical condition. Dominant hand damaged.
4) Conversion kits using any of the above
Since the aren't integral to the weapon, disabling or removing them will likely be very easy.
5) Source code of the firearm
You want to put a binary blob on my firearm? Firstly, no. Secondly I envision some interesting "unintended acceleration"-like lawsuits. (That reportedly cost Toyota about $1000 per line of source code)
Addressing all of these issues will take a long time and will be really expensive. And I'll bet the market will then end up with a DRM'd firearm that the owner won't really own. Maybe even connected to the internet, with a GPS and geographic fencing. There is a final addition which I would actually support: a screen and loudspeaker. Because then you can force the owner to view educational material on firearm safety, anger management, conflict resolution, the results of employing violence, responsible drug use and where they can find counselling w.r....
A couple of comments on your comment, at the 2012 SHOT Show there were people showing biometric trigger locks and gun safes, last year the iP1[1] (RFID) was there. At that time both Sig and Glock indicated they would have something similar this year. We'll know in a bit more than 20 days :-).
If I understood the essence of your argument, it was that this is a hard problem that gun manufacturers are not investing in solving, and under investing will only yield ineffective solutions.
I can see the point, I'll wait to see what people are showing next month before I cede my prediction though :-)
[1] http://fortune.com/2015/04/22/smart-guns-theyre-ready-are-we...
Now, maybe it would make more sense for the assassination of carefully guarded targets. But it will probably be a few more years before that technology is more reliable than strapping a vest containing c4 on someone.
That works when you are in 'gun range' of their cover and can shoot at them. But if you're pinned down? What about the enemies rear area where they are staging? How quickly can you set up and tear down a mortar emplacement?
Giving the soldier on the ground the ability to send a round exactly where they want to, behind cover, even inside a building. And I think you've created a very strong force multiplier. Given the the current state of the art in having locally operated visual recognition on the drone itself (http://www.csail.mit.edu/drone_flies_through_forest_at_30_mp...) to either avoid obstacles or target specific objects makes the weapon even more effective.
In my opinion when these things make their debut they will be as disruptive as the machine gun.
I am surprised too that NATO troops don't use smaller drones to check for snipers and ambushes.
2. Google becomes highest valued tech conpany before Apple
1) React.js starts to lose popularity due to it's ultra complex tooling ecosystem. People want to feel like what they learn will be still be useful at their next job. The React ecosystem doesn't provide that; the React ecosystem tends to burn people out.
2) Smaller frameworks like Vue or Riot takes the spotlight.
3) Angular 2 is a hit thanks to its "batteries included" design, which will appeal to React burnouts.
4) A new JavaScript rendering-based framework will come out and become a hit. Something like Turbolinks or Glimmer, except it doesn't break jQuery or sandboxes you to an ecosystem.
2) The majority of the population still does not know / appreciate what it takes to make a website
3) Computer science grads are pissed because they can't find jobs despite constantly reading news about how there is a shortage of programmer jobs
4) Native iOS / Android is still dominant, Javascript hybrid apps are only used by technically advanced companies
5) JavaScript loses popularity, MEAN stack loses popularity but alternatives aren't appearing, just more criticism of the javascript frameworks
I think, there will be alternatives. WebAssembly will let developers write client-side frameworks in C/C++, C#, Go, etc. Very likely, JS will start losing its popularity as the selling point "works both on client and server" will work for any language. However, I don't believe we can say goodbye to JS. At least, not in 2016.
1a) Release of the first non major brand (JD, Case etc) autonomous machine.
2) Drone see uptake in precision spraying applications. Although this will be on small farms. Broad acre will still be too hard for a while yet.
3) NDVI becomes one of the most commonly used inputs pre seeding and for nutrient applications.
2. Multiple new operating system projects announced, targeted for unikernel VM deployment
3. Emerging AI techniques commercialized for creative applications, e.g. a new wave of selfie apps, professional art or music tools
4. Distributed apps using blockchain protocols start making small ripples
5. Software and devices that successfully bridge the mobile/desktop gap are demonstrated
6. VR/AR devices ship, but demand remains modest and mostly in professional niches
I would have said something about the economy and finance too, but this is a tech predictions thread. So I'll go for a tech-economic one:
7. Trends turn against one or more of the current leading social networks, as a bold newcomer finds an opening
8. Bubble mania in the Valley peaks and shifts towards panic as key macro indicators start sagging
And I hope (7) Twitter's death will be quick, and not painfully slow.
Also if we do get Swift it will be for the same reason we got <canvas>. Apple wanted to make dashboard, so they put an extension in webkit, people started using it on the open web and...
Though, it's better to stay with one language, JavaScript - it's a good language, and everyone should learn it first (I would argue most simply to don't understand Javascript and use only a fraction of it, in an non-optimal way). Fragmentation is bad for the web. I would already consider CoffeeScript/TypeScript/Dart/etc that compile to JS bad because its fragmenting the developer base and it makes it harder to reuse code. Have you ever visited a GitHub page only to find out that a lovely open source project is coded in some obscure language that compiles to Javascript.
2) There will be considerable progress made on the Winograd Schema challenge.
3) The (unconstrained) Turing test won't be "officially" passed until October 2017 though (it will actually be passed in July or August, but the news won't leak until October)[1].
4) Some people will continue to insist that Deep Learning is nothing different to what was being done in the 1990s. At some point someone will get frustrated enough with this to blog everything that is different now.
[1] Specific enough prediction?
1) Ad blocker usage doubles on the desktop in Western Europe and the US by the end of the year.
2) Websites locking out users of ad blockers becomes routine, rather than exceptional.
3) There will be at least one successful legislative attempt to outlaw ad blocking, in Europe.
4) Long shot: Apple ships a minor (+0.1) update to iOS with ad blocking enabled by default. (Blackberry might do the same, but it would not be as consequential as Apple doing it.)
Bookmark so you can all laugh at how wrong I was!
I hadn't considered Australia, but you are right, that would be a good candidate as well.
1) Analytics will move back to the server, either with proxies providing log processing or new tools to provide log processing.
2) Google Analytics will react to this by providing/supporting plug-ins for most major CMS and web app software for server-side reporting
- A Tinder not for hookups (perhaps even for friends - see below)
- A small-scale Meetup.com for exchanging knowledge
- Time banks take off
1) FinTech will start to shine through, the first consumer banks built using modern tech will open to customers in Europe (London FinTech scene is strong, Zurich has ex-Googlers and strong finance, or Frankfurt but they are currently trailing - my bet is on London). These banks will experience very strong growth, the question really is: Will they go with it or go for acquisition
2) The "Family Plan" will emerge as a new sales market in most of the established consumer products, with Dropbox, Google, and others all building strong offerings for managing the product use and sharing of a group of users. This is ground-work for centralising both "family" and "home", and lays the foundation for the command and control of IoT over the next few years - it is how the big established players stay in the game.
Neither of those predictions is quite there, the leading new bank is Mondo https://getmondo.co.uk/ but it's in beta, is currently iPhone only which limits adoption in the UK, and the benefits of the new tech hasn't yet been fully realised.
And I'm not yet seeing the "family plan" head this way but I'd be surprised if the penny doesn't drop somewhere and this be the path taken. Control of the home is control of the family, and vice versa... the family want tools for this better than the ones they have today.
An entirely different thing I'd like to see exist but do not know of anyone working on at all:
3) A dating app that acknowledges the hook-up culture that seems to be growing in the millenial generation of users, and instead of hiding that under the carpet uses it's data to encourage responsible tracking of STDs and other risks.
That one is inspired by news this morning on Gonorrhoea becoming immune to antibiotics http://www.bbc.co.uk/news/health-35153794 . The app should function similar to whatever the core function of Tinder is, but allow tracking of whom you've done what with for the sole purpose of allowing notifications of STDs to be quickly disseminated to people who may be exposed.
It may be infeasible, I don't know what the willingness of those who do hook-ups is to track risk-related activity and test results is. The core idea is "disease tracking in social networking and dating apps".
I like that they have one of the things I anticipate will emerge "Instantly switch your debit card between accounts".
I hope to see features that allow the management of accounts by splitting an account into virtual buckets, allowing transactions from frequent providers to be tagged/labelled in some way.
Or even to map retailers/companies such that "transactions from them are bills and are always associated with this account", even though you only carry the one debit card it will act intelligently as simply a payment card and you would have control over where the money actually comes from.
For example:
* When travelling, you could use a mobile app to receive instant notification of every payment made, which helps you manage your fraud risk when abroad (temporarily block the card if you see something suspect, whilst allowing you to whitelist for specific transactions so your card actually still works).
* You could have control over the "routing" of your finances... designate payments as bills and have them come out of a different account than the one you use for disposable income.
* You could split payments, go for a meal and one person pays and then agree to split and the payments are auto-managed amongst those who were there (similar to Ubers split fare)
* You could have better budgeting and management of finances, better reporting, as the banks could run the equivalent of "Google ID" (Google have this for phone number recognition) for payments so that transactions are automatically tagged and managed.
* You could have dynamic control of spending limits, set alerts and controls that follow your normal pattern to really increase fraud prevention measures, whilst your app would allow you instant control when you need to make purchases outside of that pattern.
Really it's about putting consumers in control of their money in all ways.
Banks are horribly slow, and the slowness removes all control.
Edit: I should add, there is a driving force behind this in Europe: Contactless card payments.
In Europe we've had contactless "touch" payments for several years, they are ubiquitous and can be used for any payment below £30. They are now so common that most people I know carry no cash at all, they simply touch their debit cards and the transaction is done in seconds.
This has led to very noisy bank accounts, where a bank account used to have fewer than 50 transactions per month: bills + cash withdrawals, the bank accounts now have almost every transaction: a stick of chewing gum, a sandwich from Pret, a coffee in the morning.
The greater the noise, the more speed and control are real benefits to the consumer.
But the key thing is speed. Instant notification and control. Mint.com is just a layer over it at the end, rather than real-time control.
2. Major browsers will start experimenting with warning the users that they are visiting a site that does not support HTTPS (although I think that they won't be adding that to stable versions until 2017).
3. Google+ is going away for good.
4. European Union makes the progress in uniting as a single market to fight against geo-blocking.
5. Self-driving cars are still not ready to be commercially available.
The tools make it really easy to put together cool projects for 2016.
2. Digital rights weakened all around.
3. Analog media makes a comeback (books, art, film, letters, photos, vinyl, etc.)
4. Phones become uncool.
5. Donald Trump wins the election. (Democracy is a technology, or at least it is in Civilization.)
2. Social networks withdraw more APIs to protect not only privacy but their own business interests
3. Privacy becomes more annoying to lawmakers
4. Lawmakers in the US vote in more surveillance powers to force social networks to comply with their information demands
5. Lawmakers in the EU create new laws to constrain their citizens' data to EU countries
6. Distributed social networks pick up traction in both the EU and US in response to #3 and #4. One will be promoted by a major player, probably Google as a play against Facebook.
1. Development for proprietary mobile platforms (such as iOS/ObjectiveC/Swift and Android/Java) will drop in popularity.
2. Continuing increase in client-side JavaScript/HTML5 popularity. The dominant platform will be ES6/React not TypeScript/Angular2.
3. The "App Store" will start to loose its relevancy.
4. Mobile (vs. web) usage continues to grow.
And for business:
5. Social networks coming out of Asia (China, Japan, Korea, SEA - such as WeChat or Line) becoming globally popular.
6. Crash/collapse/accounting fraud scandal in a major internet company. (No idea for specifics but look at the ride subsidies of uber for an example of something that can go wrong).
7. Xiaomi starts selling in USA and Europe with considerable success.
My prediction is a better alternative to JS for client-side development. WebAssembly, I'm pointing at you. And as a result, a lot of new client-side frameworks.