Amazon ad revenue alone ($2.2 billion) for Q2 was already around 15% of Facebook's total Q2 revenue ($13 billion). That is nuts, if they have any room to grow that further, since ads isn't even Amazon's specialty.
For reference, Snapchat did $290 million in the same timeframe, Twitter $660 million.
Amazon has one of the perfect data mixes - like Google, it has all the infrastructure for true intent based advertising (sponsored listings on result pages, based on specific searches or product info), as well as comprehensive end-to-end customer journey information for your more passive advertising as well.
The latter will probably earn them decent scratch, even if it's perhaps not long-term sustainable with data issues and adblocking (Facebook can work for passive advertising because the user intent is snacks of procrastination, passive ads can be snacks of procrastination too if done well, though will always be low rent). The former though, now they're in pretty much every vertical, is where the winning happens.
As in the article, even if companies don't sell through Amazon, there'll still be an Amazon Tax if they want to get into the search results like with AdWords.
I wouldn't be at all surprised to see Amazon leapfrog Facebook, once their attribution pixel gets some traction in the wild.
> For reference, Snapchat did $290 million in the same timeframe, Twitter $660 million.
Just goes to show how much potential Amazon has in the Ad Market. Does anyone else think Amazon could overtake one or both of Amazon's biggest competitors in the market (Facebook or Google)?
Amazon is most like Facebook in that the ads are only displayed on its own property rather than all across the whole internet on any site that displays 3rd party ads. That, combined with the fact that Amazon also has a ton of data on its users just like Facebook makes them very alike (not that Google doesn't as well).
What separates Google from Amazon is that their paid ads can be displayed on any 3rd party site that they don't own, through DoubleClick. And of course the whole search engine monopoly makes that impenetrable for now.
Of course, it's all up for grabs, Amazon could go out and buy an ad network like Mediavine and go head-to-head with DoubleClick.
We've been displaying Amazon CPM display ads on a site for over 2 years now. This is an ad network for affiliates, but paid by CPM. It's mostly a mix of large brand adverts and Amazon remarketing ads (normally for products you've looked at).
The service is about to be shutdown though at the end of this month, and their suggestion is that we apply to their larger header bidding network (Unified Ad Marketplace), but not sure we'll qualify for that.
Facebook has a display network similar to DoubleClick, called Audience Network. In my experience it pays publishers a lot better than DoubleClick. For advertisers, they're able to target on more demographic categories (for now, the CA scandal has done a lot to reduce targeting options).
A significant chunk of that ~4 Billion or ~5% of total global online advertising is because Google also owns YouTube, which captures a large chunk total advertising spending independent of search.
Twitter also has an ad product called "Twitter Display" in which your promoted tweets get reformatted into display ads and shown in other ad-supported mobile apps. It was some of the cheapest, but worst-performing, traffic we've ever driven through a paid channel.
I've seen lots of money wasted on blogspam websites within GDN (non-retargeting) -- there's a need to be very hands on if you don't want to waste money across all display tactics, in my opinion
More importantly, almost all of Amazons searches are commercial in nature, whereas I think for Google it is only about 30%, the other categories being navigational and educational. This means that their data is much more monetisable.
An analyst with Piper Jaffray recently projected their ad business will rapidly exceed AWS when it comes to being their profit center. By 2021 sometime they're forecasting $16b in operating income just for ads. That would make Amazon's ad business one of the few dozen most profitable businesses anywhere.
> The analyst estimates Amazon's ad business operating income will grow to $16 billion in 2021 versus $15 billion for Amazon Web Services that year. He believes Amazon already has product search market share "well above" 50 percent. Olson noted the "Other" segment, where Amazon's ad business resides, grew sales 72 percent in the second quarter.
Isn't that like an immutable problem and what their business model is based on? Push employees to the limits to have very high performance for a short time and then throw them out when they burn out?
The problem with Amazon is that it is a race to the Absolute lowest possible price point. Now don't get me wrong, I'm not saying that boatloads of vendors don't make money advertising there (and people on Amazon are for Sure in the market), but wouldn't you rather get people to your own website via say YouTube, which is only going to get more and more popular, and then convince them to buy from you Now... I would if I was currently selling online. It simply makes more "cents."
Problem is it's much harder to get sales elsewhere. On Amazon, you can sell a $25 item and maybe spend $5-10 in ads per sale. On your own site, you might pay $25 per sale, which means you need a compelling offer at $70+ or so. This is because conversion is worse and bids are only going up as people get better funnels and can afford to throw a ton of cash on ads.
I have to say your comment sounds like an opinion, that may or may not be true; furthermore you are assuming there is $5-10 profit margin on a $25 item on Amazon... this does not seem very likely or rather an exception not the norm in the cutthroat and competitive selling arena that is Amazon.
How about AdWords; can Amazon get 50 bucks a click on a search? Nope they can't, but Google does. I don't see Amazon dethroning Google anytime soon for ad revenue. They could, but I think they'd need a high end site that caters to quality and not peeps looking to save $1.96 on a set of Tupperware.
I'm in the industry, I hear people talk about ads all the time. The $25 per action is from a talk I was at 2 weeks ago from someone doing 8 figures a year on shopify and a ton on amazon as well who made this point specifically.
Generally private label on Amazon has high margins but high competition, in the good niches at least. Everything eventually dies and gets commoditized. People spend 10-30% ACoS all day long, people give away hundreds or thousands of units. Typically people will buy amazon ads at breakeven or worse in order to gain sales rank to get more organic sales.
>can Amazon get 50 bucks a click on a search?
Google is only getting that on cars or enterprise software or other things with high purchase prices. If Amazon sold cars, I'm sure they'd be able to get people to pay 50 bucks a click.
> Typically people will buy amazon ads at breakeven or worse in order to gain sales rank to get more organic sales.
This is a very important point. I sell on Amazon and I run ads at break even profit to get more organic reviews and higher sales rank. Much more retained value compared to paid search engine ads in ecommerce
Christ I wish they could state that. When you are reporting companies like Apple, Facebook and Google which has an worldwide operation, we expect those figures to be world wide and not a specific market.
Edit: I just look up the worldwide total for Digital Ads are $228.44 billion in 2017. Total Ads market worldwide including TV, Print Ads etc are 500+ Billion per year.
Seeing ads in Amazon search results has substantially reduced my perceived trust in Amazon brand, since as a buyer I no longer feel like Amazon is on my side. I feel like I am Amazon's product. That it's me, who is for sale.
Probably a bigger issue for Amazon is that a large part of the ad market is competitors or potential competitors. Walmart, Safeway, and Azure won't be advertising on Amazon.
Does P&G want Amazon to control more of the retail market? Not if it has any brains. Amazon has it's own branded home speaker systems and will not sell Google's. If Amazon brought out a shampoo brand, would it sell head and shoulders?
Why would other big retailers pay for ads on Amazon's website? Brands are the ones who pay and already do. P&G already places a ton of advertising on Amazon for its various products. Search 'Tide' and look at the headline ad spot.
Amazon makes plenty of their own stuff, under Amazon basics, like 70 other brands, plus whatever they got with Whole Foods. Aside from the Kindle shenanigans, they haven't blocked competing brands.
They will when the counterfeit problem hits the mainstream consciousness. There's a limit to how many "genuine" Apple power adapters can self-combust before it goes really viral.
That's when Amazon will need to draw on their deep wells of goodwill, but they're squandering them today by cozying up to advertisers.
I think every company cares about its customer loyalty, and damaging the trust in your brand might leave you with the customers who use you only because you are the cheapest. This would quickly make your brand worthless.
In my honest opinion, customer loyalty is a thing of the past. I am in no way loyal to Google, Facebook, Amazon, or Apple. However, they continue to make the best products, hence I do business with them.
I've also been seeing Amazon reviews in AdChoices square-shaped banner ads. I'm not sure Amazon is facilitating this but it makes it seem like the reviews are there just to get people to buy things, and the reviews are going to keep getting more and more fake like Uber, AirBnb, and Yelp.
This is my bear case for Google. Google offers the least precise targeting compared to Facebook and Amazon.
Facebook knows more about its users because users tell it about themselves and Amazon knows more about purchasing habits of its users because of thier purchase and search history.
If I’m searching for something with the intention on buying it, I am going to go to the specific site - like Priceline, hotels.com, Zillow, etc.
Besides, ad blockers don’t work with Facebook or Amazon embedded ads.
Google has a lot more data than Facebook and Amazon combined. The targeting is not just based on your current search query. It forms a model of your behavior using your entire search history and browsing history which contains significantly more personal details about you and isn't limited to what you're willing to share in the public eye of Facebook. They also track your browsing history through every site you visit and Android app you use that serves Google ads (possibly more ecommerce than Amazon globally). They also track your location and travel patterns with Google maps and Android. All of the videos that you watch on Youtube (2nd most visited website in the world, above Facebook) tell a lot about you too. At one point I think data from your emails in Gmail was used but they stopped doing that.
True, but it seems Google has trouble refining all the raw data it mines. My go-to for passive entertainment is Youtube, and finding something new I like on that platform is like pulling teeth. I don't think anyone important at Alphabet really realizes how bad their overfitting problem is. A major reason why I sold off my investment in them this spring.
All true, but as a potential advertiser you don't really get to use all that targeting when placing ads. Although to be fair that's mostly based on my experience with AdWords.
On Facebook I can advertise to people in a well-defined geographical area, who are interested in a scarily well-defined set of things.
On Google it's more of a just-in-time approach; catching people who are searching for something in a certain place etc.
My business doesn't lend itself to Amazon ads, but I can totally see how it'd generate an ROI far exceeding Google or Facebook for product businesses that sell on Amazon.
> On Google it's more of a just-in-time approach; catching people who are searching for something in a certain place etc.
Adwords location targeting options have been around some time with 3 targeting options for location: http://prntscr.com/kq9coa
And my advice when using this - Adwords defaults to "People in, or who show interest in, your targeted locations" but I would recommend "People in your targeted locations" as the first seems to be a very loose definition by Google and in my experience brings in dud traffic.
Location settings are one of the common set-up mistakes when I see client historic accounts. And there's some good tricks with combining these settings for business that need physical proximity that can make them very effective and flexible.
>as a potential advertiser you don't really get to use all that targeting when placing ads.
you don't need to. since google is interested in providing you cheap relevant clicks, it will deliver them to you without exposing all the targetings. all the data available wil be utilized in choosing the right audience anyway.
This type of "trust us, we know what we're doing" targeting is typically less cost effective for advertisers. The fact it works at all is a technical achievement to be sure, but as far as Google goes I can vouch for the OP.
Well, if you are doing direct response advertising, you can mark (with utm source) and measure any kind of performance metric (CPA, average purchase price, whatever you can come up with), and if their clicks are not cost effective, you can stop buying them.
And if you do branding... branding is hard to measure anyway.
Note -- I'm not affiliated with google in any way, I just think their offering makes a lot of sense.
If you're selling a product that's available nationally and appeals to a reasonably big demographic, you might not care much about targeting.
But if your business only supplies weddings in New York, then impressions/clicks by people who aren't planning a wedding in New York are largely worthless.
But the initial intent still has to be there. If I sell shampoo noone's gonna see my shampoo PPC ad just because they email a friend to say they've got dry hair. They have to search for a shampoo-related keyword I'm targeting.
You can definitely do the same with Google, but yes more targeting is available as you spend more. Companies spending 8 figures with DBM/AdX get direct access to the bid stream to do whatever targeting they want.
Don't forget that Facebook also knows your browsing habits thanks to Like buttons, retargeting pixels and Facebook Connect integrations all over the web.
Absolutely not the case. The Google machine includes:
Google Analytics, Adwords/Adsense, Doubleclick network, Android and store, Chrome Browser, Chrome OS, Maps/Streetview, Gmail, G-Suite, Youtube, Google Play, Google Fiber, Google Fi, Public DNS, all the various websites with 1st party cookies, connections to Salesforce CRM, and the recently announced Mastercard purchase data.
Facebook and Amazon are nowhere close this level of exposure from hardware devices to internet protocols to the biggest websites around, and anyone with experience in adtech will tell you the same thing. Also adblockers can trivially block any network request or DOM component on a page, including Facebook and Amazon ads.
>We don't correlate or combine information from our temporary or permanent logs with any personal information that you have provided Google for other services.
They still get aggregate data that can be valuable. All the site visits that don't go through google search will still trigger a DNS lookup. Including geolocation.
All of that data is still just based on Google trying to “infer” interests which is not as accurate as people telling them about themselves. For instance, Facebook knows your relationship status as soon as it changes because most people tell them. Facebook knows where people like to eat because people check in. They know which events you are going to that are advertised on FB because you tell them “that you are interested”.
Do you know how many people check in every single time they work out?
People don’t use Google Maps when they know where they are going, but they will check in.
As far as adblockers blocking contents “in the DOM”, that doesn’t help if you’re using the FB app - one of the most popular apps in the store and most people aren’t going to install network based ad blockers.
Only geeks use Google’s DNS. Most people use whatever DNS the DHCP settings on thier router tell them. If you are using cellular data can you even change your DNS?
This means people tell it exactly what they want. Who cares what your relationship status is when I can see you type in "wedding ring" into a search box? Who cares where you like to eat in the past when you searched for "chinese food" on maps, and potentially even clicked on an ad for a restaurant?
This is called "intent" and it's incredibly powerful because you can advertise for exactly what a person is actually looking for at that moment. Inference (like Facebook likes and profile details) will never be as good for real-time targeting.
It’s not about taking general search from Google. It’s about picking off high value search.
Your random searches are not that valuable from a monetization standpoint. Searches that lead people to buy things are. Those searches are easy to pick off on a category by category basis. I had no idea that Google actually took up most of the screen on mobile with ads for products for sell until this thread. I’ve had Ad blocking on for just that long.
So easy that it still hasn't been done yet? What's everyone waiting for then with 10s of billions in profit available?
In fact it's the opposite happening with Google now pushing into searches for flights and hotels with far better UX, even though those verticals were dominated by other vendors.
You are being naive here, the edge is data size, what made Amazon stuck out is they have shopping specific data, that's what matters, so too will those specific apps.
I've been in adtech for 10+ years so I'll take my experience and connections over your anonymous comments.
Nobody cares about some app that gets 0.001% of the usage and probably doesn't even have any strong identity link. "Shopping data" isn't magical nor is it of interest to every advertiser. Purchase data is also already processed by specific companies with their own properties or in aggregate by many ad networks. And, in case you missed my first comment, Google recently closed a deal to get purchase data directly from Mastercard which is far more than any single retailer could provide.
If you actually have an example of a specific app that is delivering much greater ad performance then perhaps you should share it.
“which is not as accurate as people telling them about themselves.”
I’d argue Google’s data is 100x more accurate of who I am as a person than the data I provide FB. As a theoretical example, based on my search history, Google knows I have hemorrhoids, am paranoid about pancreatic cancer, looking for a SMS API provider, am a borderline gambler (because i check betting odds every day), and pondering the benefits of a divorce.
FB on the other hand has all the “shallow” information. It knows I like Game of Thrones, NBA basketball and paintball. It knows I like to eat ramen. It doesnt know all the other stuff because I dont tell my friends I have hemorrhoids. I dont tell my friends I might be getting a divorce.
Does FB know the needs/wants/fears of me as a person compared to google? I think not.
Facebook also has Messenger, which to my knowledge and from my experience is far more prominent than Google Hangouts for "private" digital communication. Presumably they mine those conversations for ads as well.
I'm not sure how important it is but I think it's also important to note that even if people are telling Facebook lots of things about themselves that there is no guarantee that info is 100% accurate.
Fair, but on the whole I'd be more willing to bet based on a willfull Facebook action than infer from an indirect Google Search action.
Data mining is incredibly insightful in aggregate, but my searching for "Cheetos" is a difficult path to (a) my being hungry for snacks, (b) my wanting to see the mascot, (c) seeing if they have a new flavor, etc. etc.
Also, if you own Android, no need for Google maps. They are already tracking your mobile location. Scanning WiFi in the background. They know you are having dinner at some hotel because WiFi found its network while scanning. It doesn't matter if your WiFi is off.
Reviewing for Google map doesn’t give you the same social media cred of “look how exciting my life is and how great my relationship is!” that checking in on FB gives you. Besides it’s a lot more work.
They know the vicinty. If there are a lot of restaurants or other businesses, they don’t know exactly which business I was at. If I have an iPhone. They don’t even know that unless it was some place I needed directions to.
There is a huge difference. Google tracks location data on Android users all of the time. Google can only get location data on iOS users if they are actively using Google Maps. You don’t need a maps app if you know where you are going.
You also don’t have to sign in to Google Maps or connect it with “you” on iPhone. It’s sandboxed unless you opt in to allowing it to use your location all the time.
On iOS, your location isn’t available to Google though. iOS has three per app settings for location permissions for apps - never, while using, and always. Google can only get location data if you are actively using Google Maps.
You're stuck on Google Maps and checkins but that is not the only thing that people use, nor is GPS necessary. Anyone sitting down at a restaurant and using the internet on their device via any app or website will shed location details that Google and other networks can pickup.
How does Google pick up location data on an iOS device using Safari browsing Facebook? Can a site with Google analytics gather the logged in Google account name? The upcoming iOS 12 version of Safari is going to make it even harder to do browser fingerprinting.
Networks have locations. Cell towers, wifi hotspots, corporate routers, etc. Not hard to get a location as long as there is network traffic.
Google analytics does know the same user across sites and devices but this is of course not exposed to the site, it's used by Google for ad targeting.
Browser fingerprinting is not necessary because people log in (to apps and sites) and will never be possible to fully reduce unless Apple can somehow make identical computer chips down to the atom.
Apple's cookie war also only helps Facebook and Google who maintain active 1st party cookies with every user, and meanwhile hurts all the independent networks and publishers who have to get people to keep logging in. This is why it helps to actually have some adtech expertise and industry cooperation instead of creating silly browser rules and turning Safari into the new IE.
Networks have locations. Cell towers, wifi hotspots, corporate routers, etc. Not hard to get a location as long as there is network traffic.
How does that help unless they are sharing the data with Google?
Apple's cookie war also only helps Facebook and Google who maintain active 1st party cookies with every user, and meanwhile hurts all the independent networks and publishers who have to get people to keep logging in. This is why it helps to actually have some adtech expertise and industry cooperation instead of creating silly browser rules and turning Safari into the new IE.
If you haven’t noticed, Apple isn’t trying to help the adtech industry. Apple has a built in framework that allows third parties to submit ad blocking rules into Safari and (some) embedded webviews.
I didn’t know how bad Google’s search results were with most of the page taken over by shopping results on mobile until I started loading results with content blocking turned off to post on this topic.
Not if I’m on an iPhone. They know the vicinity I’m in but they don’t know exactly which store I’m at if it is an area with a lot of stores nor do they know that I’m with my friends. I probably added them to the check in and/or tagged them in a picture.
You do know that Apple only tries to stop things like this for everyone else than themselves, right? I wouldn't trust them any more than Google. At least you can use an android phone without Google apps if you want to. Good luck with using an iPhone without sending your data to Apple or trying to remove all Apple's stuff (step one: iOS).
Unless Apple is lying and third party research says they aren’t, they only store location data locally on the device and have an entire framework where they anonymize location data sent back for improving their service - differential privacy.
For the really paranoid, you can completely turn off most sharing with Apple, disable iCloud backups and backup to your PC with iTunes.
> Google offers the least precise targeting compared to Facebook and Amazon.
In my experience, Amazon has _incredibly poor_ targeting. Search and place an order for cat food, and the next day you'll get an email with:
> Based on your recent activity, we thought you might be interested in this.
>
> ... cat food, the exact same brand I bought already ...
Not only that, their targeting doesn't seem to be based on my past purchases nor my frequent purchases, but almost solely based on what I have searched for recently.
They don't seem to use the _massive_ amount of data they have on my purchase history _at all_.
I bought a TV on Amazon a while back, they knew I looked it up and they knew I paid money for it and they knew I received it and I even left a review so they knew I used it. I got ads to buy that same TV that they knew I already bought for months after I bought it.
The most predictable purchase for someone who has purchased one TV and reviewed it positively is the exact same TV for another room.
They are just following the money, it might not make sense to an individual, but product recommendations are optimised for sales based on historical data.
That's a really good point that I hadn't considered before: sometimes the ad that maximizes the expected return won't be the ad that is most accurately targeted. If an ad for a $10 HDMI cable for your new TV snags 10% of consumers, that's a well-targeted ad, but if 1/200 consumers will purchase another $500 TV if you remind them about it occasionally for the next six months, that's the ad it makes sense to run.
That is true... I bought a flash drive on Amazon with the intention of buying more of the same flash drive if I liked the performance of the one I bought. And when I liked the first one, I bought four more.
A TV seems excessive though, but maybe that's just me. I mean it was a $700 TV, how much money does Amazon think I have to buy two of them?
> Google’s U.S. ad revenue is expected to rise $5 billion this year to $39.9 billion, while Facebook could rise $3 billion this year to $21 billion, according to eMarketer.
Adwords is just so effective to find your customers. You won't beat `buy new headphones` type traffic on Amazon or Facebook.
Personally, Amazon is where I go to check prices after I've decided what it is I'm buying. I've never gone to Amazon without knowing exactly what I wanted.
Your experience doesn’t jibe with market realities. Merchants play all sort of games to be the number one listed search result when you are searching for something generic like “headphones”.
Who actually searches for such a short tail term and expects good results? Does anyone ever google the term “cars” when they want to buy a car? More likely it’s a term like “8 seat SUV” or something more focused. Maybe headphones are different though..
And the first non advertising link is car.com where I can search for specific cars. Now Google has lost a customer and car.com has gained one. Now I can advertise on car.com and get a much better targeted advertising.
I never start a product search on Amazon unless I know the exact product I want. With headphones, I’d be likely to care about reviews in audiophile magazines long before I’d enter the noise of Amazon. I don’t generally use Amazon for product discovery, but except perhaps for paper books and I want related titles to something.
As long as your happy paying a premium in many cases for consistency. Certainly a lot of the marketplace stuff can be found outside of Amazon for cheaper as the merchant has an extra 15% to play with gong direct.
Most people are happy to pay a premium for the convenience - as evidenced by all of the third party merchants who sell on Amazon even though setting up a website to sell merchandise is trivial even for someone who knows nothing about websites.
I just searched for “buy new headphones” on Google with my ad blocker turned off.
It showed ads where I could get headphones from different web sites - where neither I nor most people have an account.
What’s more likely? That I am going to register for an account on “Under Armour” - the first site that appeared or that I’m going to go to Amazon where they already have all of my information,I’m an Amazon Prime customer and where I have one-click activated?
If Under Armour were smart they would have web based Apple Pay payments available to reduce friction.
I think the clear answer is it's not for you. Part of the reason behind selling through other channels is to capture people like you: you (and me for most things) would rather shop through Amazon than a retailer.
Retail vs. online is a prime example of channel selling. Some people prefer to try on things or are so brand loyal that they're blind to other retailers.
Yup. Amazon also has the advantage of dark UI patterns; they put sponsored listings inline with normal product results in the exact same style. I'm the (probably atypical) sort of person who will deliberately not click on a sponsored listing even if it seems the most relevant, and I still regularly click sponsored items on Amazon by accident. I back out if I notice this, but I'm sure I've bought a few products this way. These seem like very effective ads to me.
Arguing that an incumbent business is impossible to disrupt in a large market seems like an odd thing to do on the forum of YC, an accelerator that pretty much specialises in supporting startups that disrupt large markets and topple incumbent businesses.
Not that Amazon is a startup exactly, but they certainly act like one a lot of the time.
There's something to be said for the expectation vs reality PR creep factor.
People searching on Google (or checking in on friends and family on Facebook) or browsing the Internet don't expect their history to be mined for advertisements.
People browsing on Amazon have already signaled their intent to shop.
And if Facebook is advertising tech related merchandise based on me putting that I am a developer in my profile, it’s a straight line from A to B that wouldn’t creep me out.
If we're talking about data I think Google wins. Email is the activity stream of all your online purchases. Google analytics has the users' clickstream data. Google messenger and google+ gets the relationships the user has. Youtube gets entertainment data. Search is what is relevant to the custom at the moment of time.
If I was Amazon, i'd look at buying mint from intuit to get all the banking data from users.
When I perform a google search for something with no adblocker on, 'sponsored results' show up -- google ads, on google's website. Adblockers hide these. How is this different from fb or amazon aads?
The majority of users are using FB on mobile and the FB app is one of the most popular. Ad Blockers that most people are using on mobile - either via the built in content blocking framework on iOS or by using another browser on Android, don’t work on FB.
FB uses “native ads” and “boosted content” that are indistinguishable from regular posts.
Even on desktop FB's ads are unblocked (or were when I used to have a FB account). They use a significant amount of obfuscation to make it impossible to use typical adblock techniques (url filter, css selectors, simple heuristics) without breaking the rest of the site. So far as I can tell most of the filter maintainers leave Facebook's native ads alone.
Google shows you ads based on what you type in so you are directly telling them what you're looking for at that moment. That's a pretty good position to be in as an advertiser. Some people will do what you do and go right to a specific search site, but others will start that process at Google.
Amazon can do the same, but the user is probably already looking to buy, although with less breadth of category coverage. Amazon is also more of a closed ecosystem, which is less flexible for advertisers but probably better for conversion.
> If I’m searching for something with the intention on buying it, I am going to go to the specific site - like Priceline, hotels.com, Zillow, etc.
And quite often that specific site will tell Google everything they know about you because it wants to advertise on the AdSense network.
For example, if you go to Singapore Airlines website and search for a flight to Thailand but do not buy it, then open the New York App to read the daily news, you will likely see an ad for Singapore Airlines - perhaps even showing a special deal on flights to Thailand.
I look at it the other way. People keep comparing Amazon and Facebook to YouTube, But I don't understand why. YouTube is a single channel and is very different from others because they serve video ads.
Google (search) on the other hand, seems to be a much closer competitor to Amazon because lots of people search for products there and then use Google's product search to find the product they want from many different retailers. It's sort of Amazon vs. the field. Do you want to limit your search to Amazon and its platform or expand your search to every online retailer who doesn't sell on Amazon? I've found Google product search to be extremely useful in finding the right combination of trusted retailer and lowest price (including sales tax and shipping). Add to this a list of highly-ranked reviews or sites related to the product I'm searching for directly beneath the list of products, and I usually find Google a better shopping experience.
> Facebook knows more about its users because users tell it about themselves
I'd be surprised if Google doesn't know more about its users than Facebook. Google (theoretically) has access to its users' email, physical locations, browsing history, and more.
And search ads remain dominant for one time, high value purchases (think insurance or a mortgage).
And Facebook knows where I actually go based on check ins. My hobbies and interest, the movies I watched, etc. People tell Facebook what they are doing. Besides people don’t use email for personal communication as much as they use apps like Messenger and WhatsApp.
For high value purchases like a mortgage, wouldn’t better targeting be by advertising on a site like Zillow?
Google knows all that things too, without you having to expressly tell them.
Where have you been? You are constantly tracked if you have an android device. (not sure whether it also works for iOS with google maps on). No need to check in.
Hobbies and interest - search history.
The amount of granular information that Google could potentially have on an individual person is terrifying. The search history alone is effectively a peek into your stream of consciousness.
How does Google know my birthday, exactly how old I am, the events I said I was interested in, everytime I change jobs, my relationship status, the college I went to, where my wife and I went to see a show the other night, whether I am single, married, or somewhere in between? This is the kind of data that people tell FB about explicitly. Google only knows that have a vague interest in running. FB could have a better idea of when I go running because I might have posted about it or used one of the social apps that automatically post to FB.
Google can only “peek”, FB is having an active conversation.
Google doesn’t have the level of data for iOS devices on locations. iOS users in western countries are far more statistically valuable to advertisers than Android users.
I'll strictly speak about Google. (and assume that you are a regular internet user, who has gmail, and doens't go to great lengths to mask their activities)
If they wanted to be really invasive, they could get a lot. Not saying that they are doing it, but they can
1) Birthday - on every birthday, there are a bunch of random sites where you have an account that send you congratulations.
2) How old - beyond patterning (which can get you within a fairly narrow range), I am sure that most people have received or sent an email which states their date of birth - whether in the text, or as part of a form
3) Job changes - in any european country, you get your employment contract via both mail and email. (and btw...this means they have your salary too)
4) College you went to - if you already had gmail while at university, they have you. if you got it afterwards, they probably saw the "alumnus" emails
5) Relationship status - this is arguably trickier. There's no "hard" proof. I think FB has the edge here, though not as large as it used to be (I remember that 10 years ago everyone had their status on FB. now it's almost no one, but it's irrelevant, because they can look at your photos, and whether you mention "gf/bf/wife" etc.
6) Running - again, FB have the edge in determining how much/where etc. But google know you are running, because they have your emails from the apps to which you subscribed.
Might I ask what demographic group you are in? I personally am in the "founding generation of FB" (now a early 30s engineer in the US) and it's obvious to me that at least the people I know aren't giving data that much to FB. While plenty lurk or use messenger, quick sampling shows < 10% of my friends update more than once a month.
So to answer your questions:
> How does Google know my birthday
> exactly how old I am
I'll give you FB has exactness better specified, though I question how relevant it is. Regardless, gmail could easily figure this out by looking at inbound emails. Also linkedin public profiles.
> the events I said I was interested in
Given the lack of updates on FB, search is a much better proxy for this.
> everytime I change jobs
Email + linked in
> my relationship status
Search and email is going to be a very good proxy here, possibly better if FB updates are stale.
> where my wife and I went to see a show the other night
Maps. (And I'd guess usage of maps far outweighs people who voluntary tell FB)
> Google doesn’t have the level of data for iOS devices on locations. iOS users in western countries are far more statistically valuable to advertisers than Android users.
Yes, but number of Google-Android users, let alone users who use Google products, probably far outweighs frequent FB updaters.
Might I ask what demographic group you are in? I personally am in the "founding generation of FB" (now a early 30s engineer in the US) and it's obvious to me that at least the people I know aren't giving data that much to FB. While plenty lurk or use messenger, quick sampling shows < 10% of my friends update more than once a month.
We aren’t talking anecdotally about “people you know”. Facebook has more than 2 billion daily active users.
I'll give you FB has exactness better specified, though I question how relevant it is. Regardless, gmail could easily figure this out by looking at inbound emails. Also linkedin public profiles.
Google claims it doesn’t use email for targeting ads and how will it correlate LinkedIn profiles with Google users? I doubt that MS owned LinkedIn is going to share the data. LinkedIn only has 260 million active monthly users.
Given the lack of updates on FB, search is a much better proxy for this.
Lack of updated by your circle...not based on published numbers.
Search and email is going to be a very good proxy here, possibly better if FB updates are stale.
How many people use email in 2018 for personal communications?
> We aren’t talking anecdotally about “people you know”.
Agreed, I'm raising anecdotes as I haven't seen any data users frequently update.
Quick search shows https://zephoria.com/top-15-valuable-facebook-statistics/. If you do the math, there's about 1.3 status updates/week/active-user. That's not a lot of voluntary data being given per user and I suspect this is subject to a huge power law where median is far less than 1.
> how will it correlate LinkedIn profiles with Google users?
Profiles are public and picked up by Google search. Correlation is pretty easy to do - simple method is real name match + geo-ip.
> How many people use email in 2018 for personal communications?
I don't have solid data; do you? That said, I and almost everyone I know uses email as the primary means for long-form (i.e. anything that requires more than a sentence) personal communication.
(I feel like we're talking past each other due to the lack of data + being in different circles)
> People only use maps when they don’t know where they are going. People will check in a lot more.
Source on both pieces of data?
I'm skeptical as:
1. Maps is valuable for traffic-adjusted navigation.
2. Maps has 1B montly active users
3. Check-ins are rare. (looking at my feed - I can't find numbers)
How much personal email are people really doing in 2018 that’s not not either bulk email or business to business (meaning Google wouldn’t have visibility), etc?
> This is my bear case for Google. Google offers the least precise targeting compared to Facebook and Amazon.
No way in hell. Google knows exactly what you are "searching" for. Sure Facebook is good at "guessing" but Google doesn't even need to guess. You just tell him what you are looking for.
Sure but if you have a SaaS product then Google seems like a better fit because you can target keywords instead of humans since you don't really know if there are any similar characteristics to your buyers.
Google realized this a decade ago and knows that the search engine is becoming a relic of the past, that's why they tried to pivot to social networking/media so strongly.
I suspect they are still working on this pivot in various non google plus ways. Outside of android becoming a massive media platform I don't think much has worked. (Don't call it a media company we don't control the news we're just a platform la la la I can't hear you controls the news)
We tell google about the purchases on our site 3 times. For AdWords, GA and Trusted Store reviews. They the have plenty of data on what is likely to convert.
Facebook is dialing back the targeting you can use, it can take days to get your ad approved and I had an account for an insurance agency not be able to use the FB lead form capability because FB said it would be used by third parties. Fb might not be aligned to use their own strengths.
But what users tell facebook may be completely incorrect, and might be more a reflection of how they want others to see their profile, for example liking things which may make them seem cool.
Google on the other hand, should know your deepest hidden desires.
Facebook has terrible targeting. They might collect lots of data but they truly suck at utilizing it. It is pathetic. Google is fairly mediocre. Literally nothing in the quality of online advertising has changed since 1999.
Sidenote, adblockers work with Amazon and Facebook.
I'm curious about your experience here. In my experience Digital folks come with a lot of smoke and mirrors that crumble under standard statistical analysis.
This is missing some testing. In multichannel environments you have the potential for multiple touches, as well as baseline if your company is large enough. For example, a traditional product like cars: you have a baseline of folks who come in, and then you have (ideally addtional/higher rate of) folks that come through the digital channel, such as engaging an online salesperson. Do the digital channels generate a higher rate of purchase than the baseline channel? If so, is the lift statistically significant? These are what I mean. Revenue moving through a digital pipeline could be because people happened to fall into that bucket but were already going to purchase anyway. In other words, some of the $1000 went to non-incremental marketing. The goal is to make sure the incremental marketing exceeds the spend.
Honestly in my experience they don't. The low quality ads is a major reason I have ramped down my Facebook usage. Google is sort of ok. The targeting on every platform is still 1999 quality. There literally has been zero change to the end user despite all the data collection. At the end of the day, it's low quality advertising.
It also makes it possible to find users who are at different specific stages of the buying process. That's way more powerful than what FB or Amazon offer.
If I want to target a demographic for top of the funnel advertising, I can target a lot better with a Facebook. If I want to target people who are buying now, why not go to the source - Amazon, Zillow, Priceline, EBay, etc.?
I don't know. Google has some pretty precise targeting of intent with search terms. But I agree that it doesn't have anywhere near the wealth of user data.
Ad blockers are used by 600m devices.
40% desktop
15%. Mobile
(at least in the US)
Happy to gather sources if you'd like, I searched this up a few days ago
It's worth noting that Amazon showing ads on amazon.com is more like a "slotting fee" or "slotting allowance" than a display ad. Given that you're at amazon.com because you want to buy something, it's no surprise that online slotting fees can be lucrative. But it doesn't mean Amazon has much room to grow that beyond their overall growth.
It probably depends on the value funnel. As you get closer in rellavance to the target the ability to influence ought to be most valuable. I question if they are monetizing the ad value funnel fully.
I heard at a talk a few months ago that if you are planning to build an app to sell to retailers, you should build it on any cloud except for AWS. This is because the large retailers are uncomfortable with giving any revenue to Amazon.
I heard echos of this in the last paragraph of the post, but there's plenty of co-opetition going on as well.
Sort of. Some retailers have chosen to use other providers because they compete with Amazon. But anyone building a new app is a LONG way from that being a roadblock to success.
If it is a B2B app, maybe not that far :). But I agree, if you get to a certain size, you can move no matter where you start. I think it'd be an interesting angle displaying customer empathy though.
I don’t think AWS is literally cross-subsidizing the retail business, but it allows Amazon to keep running retail at essentially zero profit without bothering Wall Street. Even without direct subsidization, that’s a serious advantage that no one else can really hope to match.
I've been long amazed at how bad Amazon's web ads are, as if the department was authorized to get rid of money as fast as possible. Besides the times -- because of one click through to a joke product like $10K ultrapremium audio cables -- that I would get followed by weeks with absurd ads, my favorite example is getting barraged by Battlestar Galactica season 3 ads right after I bought the damn thing. OTOH, perhaps Amazon's ad/recommendation system was too smart, trying to warn me from purchasing season 4.
I understand that big data learning systems are harder to tweak than what we'd suspect from ground-level day-to-day experience. But it just seems like Amazon still has plenty of room to apply some conventional heuristics to get far more efficiency than they do now.
Those ads are mostly on Amazon's automatic remarketing systems. They only recently opened up control over the bidding and targeting of those to sellers and brands. Eventually considering how well that ad system tends to work in the right hands, I would expect it to be less intrusive and more effective over time.
Amazon's remarketing tends to be paid for by Amazon and it behaves like any remarketing system run by other companies if you don't change anything from the defaults, set a very high frequency cap, and don't pay attention to whether or not the person converts.
"Leaders start with the customer and work backwards. They work vigorously to earn and keep customer trust."
Today:
> The push by the giant online retailer means consumers — even Prime customers, who pay $119 a year for access to free shipping as well as streaming music, video and discounts — are likely to be confronted by ads in places where they didn’t exist before.
Was this really the outcome from "start with the customer and work backwards?"
> ...it [Amazon] is rapidly emerging as a major competitor to Google and Facebook.
Gotcha. We all know who "the customer" is in this market.
Seriously, completely wrong direction Amazon. Our trust is already eroding due to fake reviews, counterfeit items, and Amazon Marketplace nonsense. Racheting up ads will make us further question where your loyalties lie.
>Our trust is already eroding due to fake reviews, counterfeit items, and Amazon Marketplace nonsense. Racheting up ads will make us further question where your loyalties lie.
It doesn't matter until it matters in aggregate and affects share price. Whenever Amazon refers to "the customer" remember what they really mean is "the customer's money," which they're still getting plenty of.
I've been getting strange ads on facebook for industrial kitchen items on amazon recently... i didn't think i wanted one, but now i am considering an at home industrial orange juice presser...
Slightly different topic but what are the big ad networks doing these days to address the proliferation of malware across their networks? I've been having a lot of issues with browser hijacking lately from ads that appear to be served from legitimate ad networks (after some cursory investigation namely from Doubleclick), and typically from HN submissions.
I've attempted to report these issues on a number of occasions. But the amount of attention it gets suggests at worse a level of complicity.
> Thanks to the vast amount of data Amazon collects from its customers, it can target ads not only to basic demographics — say, women over the age of 35 — but to a more precise segment of customers who are likely to be shopping for cellphones or barbecue grills.
...and later:
> “We can reach the right consumer at the right time using their wealth of data to target,” Ms. McGurk said. “Other traditional digital platforms do not have the level of purchase data that Amazon has on their customers.”
It seems rather disingenuous to paint Google and Facebook as the bad guys for selling out their users' data for ad revenue but then go on to say that they are behind Amazon because Amazon has so much more useful customer data and is willing to sell it. It's becoming clear that the NYT is going to war with Facebook and Google (understandably so, since they killed their business).
One final point- in a time of economic expansion, which we have been in for a few years now, ad spending is going to rise. All players- Facebook, Google, Amazon, probably even the New York Times- will see better results. There are more dollars floating around the economy as consumers spend more instead of save and advertisers feel that they need to spend more across all channels just to keep up with their competitors.
Amazon knows way more about its users than Google and Facebook. It is easy to deduce a lot of information from my purchase history, my reviews, my 'helpful' clicks, my address, etc.
I can switch to DDG, Telegram, and Yahoo Mail, but I certainly will continue do shopping on Amazon - there is no easy escape from it and I wonder why?
Amazon is already an ad agency for vast number of 3rd seller while also offering fullfilment services. You can see the impact of transitioning to this business model. Large number of those “sponcered” listing is crape. A sizable 3rd party sellers are straight from China looking to dup the customers. Significant number of products that I bought past few months are branded like Duracell but completely fake.
If I run some large customer-facing services on AWS (or any service, for that matter), doesn't that provide Amazon with the same (or even better) detailed insights on customer behavior I can have?
Am I the product as well as the customer on AWS?
While Google with their cloud services has the same opportunity, this additional source of data is not available to FB.
Services you run on AWS are isolated from Amazon. It is part of their legal agreement. This is important so that businesses can store sensitive data on AWS without worrying about Amazon using this data for their own use cases.
My friend in commercial real estate said something I thought was interesting: the cost per acquisition in a mall can often beat the online CPA. I had assumed malls would have much higher CPA. It'd seem low-priced point retailers like Bonobos lend creedance to her claim.
They will try to sell access to you based on the information that they have about you, but they will generally not sell the information they have about you. (For one thing, they could only do that once...)
is there anything amazon isn't going to "set it's sights on"? will they be building and selling Amazon Houses and providing Amazon Health Insurance soon?
For reference, Snapchat did $290 million in the same timeframe, Twitter $660 million.