When a random person searches for a company, it's probably > 100:1 odds that they're interested in the company's product, not in a job at the company. That's why companies put their products on the landing page and not their job openings.

You have to understand that whenever you do anything as a consumer in modern corporatism, you are part of a numbers game. The company does not care about you, they care about the aggregate "you" that represents the average of millions of consumers. If 100x as many people search for a company hoping to buy its product, then it only becomes rational to spend engineering effort if the opportunity cost of losing the one customer who is job-searching is 100x greater the lift that can be achieved by fine-tuning the product-buying case. Retargeting would have to be pretty tapped-out for that to be the case.

I face a similar problem - as an entrepreneur, when I Google a company, it's usually because it's a potential competitor that I want to learn about. As a result, my YouTube & AdSense feeds are filled up with ads for competitors. Which is kinda handy, in a way, but probably not why those competitors are buying ads.

> When a random person searches for a company, it's probably > 100:1 odds that they're interested in the company's product, not in a job at the company. That's why companies put their products on the landing page and not their job openings.

What exactly is the feedback mechanism to discover this "fact", or refine its expressions and exceptions? This is a self-reenforcing presumption.

I'd arrived at it by estimating that the average person will end up comparison-shopping for about 100 different products or services during the time that they hold a job, but will have only one job. If you assume 2 years at a job, that's about 1/week, which seems reasonable, counting all the gifts you buy, restaurants you visit, trips you plan, service providers you look up, products you evaluate for work, etc. Wouldn't be surprised if it's even more.

For any given business trying to optimize their website, it's a lot easier than that. Just look at the traffic flows in Google Analytics. If everybody hits the landing page and immediately clicks on "Careers", maybe the focus of your landing page should be on the job opportunities available. If they don't, you're probably right to focus on the product. If a significant number of people visit "Careers" but not enough to move it to the front page, you may want to exclude visitors to it from your remarketing campaign. (A sibling comment indicates that this is trivial for both Google and Facebook ads.)

That more sites don't do this - when it takes all of about 15 seconds to diagnose in Google Analytics and a couple minutes to fix - is probably a good indication that the economic losses from this situation aren't all that significant.

> When a random person searches for a company, it's probably > 100:1 odds that they're interested in the company's product, not in a job at the company

Really? There are companies that have very few customers that pay lots of money for very specific machinery or services. Customers which might be banks, large businesses or large manufacturing companies. How many among the visitors of their websites are the decision makers that are evaluating their products and those of the competition in view of a purchase? You say 99%.I say 1 out of 1000.

There are, but again, we're talking in averages, and we're looking only at transactions where Google makes money. If a VP at a bank buys some financial software for $10M, it's great for the vendor, but Google still only makes about $10-20. If an engineer at the bank searches for software to make their job easier and clicks on the link, Google still makes that $10-20, but no sale happens (immediately). The financial software vendor has folded the costs of those clicks by non-decision makers into the lead-gen cost of their product, and I would bet that there are 100x more clicks from engineers, managers, and salespeople interested in the product than there are from VP decision makers (or people interested in working at that firm, for that matter).

And that 1 out of a 1000 person might buy a product/service worth millions which makes the marketing and acquisition costs worth it.

You're looking at this through personal assumptions and ignoring all the data when this entire industry is completely data-driven and calculated in decisions.

Of course. I'm not saying that the advertisers are wasting their money. I'm saying that it seems that both the advertisers and the ads network could have made more money with a better targeting. At least in my case, they've wasted the cost of the impression and the potential gain from showing a more relevant ad in that space. And this is probably happening millions of times a day.

It is, but the cost of better targeting gets exponential more expensive.

Also people complain about targeting and tracking and delete cookies so there's that whole situation (which really only hurts them by giving them worse ads).

> It is, but the cost of better targeting gets exponential more expensive.

Does it? I don't see why. It's not like you have to manually go through the data.

What magic data is there? As I said, there are privacy concerns which are already limiting the quality/quantity of data available and leading to more generic ads. Outside of that, it takes ever more data and the required storage, processing and algorithms to derive exactly what your intent is.

You can look at any algorithm problem and see the decreasing returns (netflix recommendations are a good example). This industry is not full of idiots, there are thousands of data scientists and researchers working on this. If it was that easy, it would already be done.

Funny how this comment, which seems perfectly reasonable to me, was downvoted twice. No explanations of course.

A few comments on this:

1) The company might be better off capping their impressions to you.

2) The company might be better off focusing on setting their remarketing pixels further down the funnel (unless you entered their funnel as part of the process).

3) How do you reasonably expect the company to treat you differently than they do actual leads? I'll assume the majority of traffic that goes to their website is interested in their services and not researching for a job interview, so unless there is some way they could reasonably figure out you're not interested and exclude you, I don't think what's happening here is that wrong. Marketing to people who already visited your site is almost certainly going to be worth more money to advertisers who are prospecting based solely on your interests.

This.

Statistically, based on limited information, re-marketing to groups of people who have already visited the site (or merchant, for offline marketing - I've done this in meat-space too) will beat the pants off marketing fresh leads.

I do agree with the impression cap comment and bumping the tag down deeper into the site / funnel, although a really strong job hunter would read the full product description (like any serious prospect)....

I don't understand why the job of figuring out how to place the ads best should be a concern of the company, instead of Google itself. Google knows basically everything about me, including where I work and what type of job I do, I'm surprised they can't do a better job at placing their customer's ads.

You have to understand how ad exchanges work (and what they are) and what intermediaries there are.

Ad Exchanges handle the actual delivery of ads. They have a massive inventory of ad slots available on a massive amount of websites and it's their job to fill those ad slots up while making as much money per ad slot as possible. Google (Doubleclick) is not the only display ad exchange [1]. There are quite a few other ones like OpenX, AppNexus, Microsoft, and so on. You can directly buy on these platforms but it's generally better to use a Demand Side Platform (DSP) because they provide a wrapper around these platforms and you can manage bidding, targeting, etc. in one place. They also offer various programatic techniques like real-time-bidding (RTB) that can end up costing you less than you'd spend directly in an exchange. Lastly, some exchanges operate in such a way that small advertisers can't utilize them. It's a human-intensive process that involves spending a lot of money, signing and committing to insertion orders, etc.

Because of this stuff, most people are now using DSP's. The DSP's buy large batches of inventory from the ad exchanges and it's their job to maximize the revenue they get for their purchased inventory. As an advertisers, I can set up my campaigns in the DSP, which might involve setting targeting options, setting up remarketing pixels on my website, etc. With RTB, I set a maximum bid amount (in CPM) in my campaign and when it comes time to serve an ad, if the ad slot matches my targeting or remarketing preferences, I compete in an auction against others and the highest bidder gets the impression.

In almost all cases, the ad exchanges are going to make more money (by means of providing value to end-advertiser) selling to people who are setting their own targeting preferences rather than trying to guess. Within the targeting options presented to advertisers, there is lots happening in the background by machines that are doing things humans can't. Ad tech is pretty sophisticated, even if it sometimes seems that it's not. Ad tech has created much more value for advertisers than it's lost when it isn't perfect.

[1] Many people may think Doubleclick is the exchange but that might be because Doubleclick is the most common ad slot you see on a webpage. However, that ad slot is a different product, Doubleclick for Publishers (DFP). You can drop in inventory for any exchange (or even hardcode an ad in cases of a direct buy) in DFP, so it's displaying ads from other exchanges, not just Doubleclick.

Maybe because they don't 'know' in the way you seem to indicate; Google's not a person, making those inferences is hard - clearly the state of the art is very much advanced from even 5-10 years ago, but how do the algorithms know that you don't also want to buy things from the company you applied to? That's not a simple inference, given a (by necessity) limited graph without human capabilities

What Google knows about you, or Facebook for the matter or any other buyer, can be used for targeting and as user features in the ML model which is used to determine the price the buyer is willing to pay for a given impression in the auction.

Most advanced buyers with actual ML in their buying algorithms do this. But ML works in statistical averages on the behavior seen across all users visiting a particular site. At that point the buying process works by figuring out the expected value of a new impression and bids that value, the expected value depends on how the advertiser values clicks or conversions or impressions, so as long as the cost of the impression is lower than the marginal value an algorithm will continue to bid, and potentially win, because it's worth it.

And you can do all the A/B tests you want and you'll see that this is actually true, capping frequency. or choosing to not show an ad because the position on the site is not great, is not a good idea, the right process is to determine a price that, all things considered, is the maximum price (proxy for value and risk) you are willing to pay to be shown in that bad slot that adds marginal value for the advertiser, and marginal value is measured however the advertiser wants.

Thanks for bringing some sanity to these comments. These sorts of articles tend to bring out those who know absolutely nothing about the space or how it works. That said, this reminds me I need to build an exclusion list for our jobs page ;)

They can absolutely figure it out, just not for less than it costs to throw the ads at you, which is very little.

They're not wasting money in showing me that ad, its cost is negligible. They're wasting the money they could have made by showing me a relevant ad. For example, there's tons of books I might be interested in buying, and they might be able to extract that information from my browsing habits. It would be an advantage for me as well, to get suggestions about things I might be interested in.

Showing me nail varnish ads when I already told google that I want to buy some is not even advertising, it's just paid search.

It's hard to infer you want to buy books from a search about nail polish, mixed with whatever else is in your history.

There is also the question of who's willing to pay for ads, and for much are they willing to pay. They have to pick from what's available. I see a lot more ads about stuff that is costly and profitable, i.e. photography gear, car rental, insurance.

Google knows where I live, what job I do, where is my office, at what time I go to work in the morning, and almost every page I visit on Wikipedia. It even knows where I parked the car the last time. Is it possible that with all this information it can still mistake me for a potential buyer of that company's product?

I think that ads tend to be about costly stuff (if that is true) because they have to discount their own low accuracy. You can make money by showing me ads for goods worth only a few dollars, if there are enough chances that I'll buy them.

> Google knows where I live, what job I do, where is my office, at what time I go to work in the morning, and almost every page I visit on Wikipedia. It even knows where I parked the car the last time. Is it possible that with all this information it can still mistake me for a potential buyer of that company's product?

Why would you expect them to be able to? Algorithms aren't mind readers. Just because a human could look at all those individual data points and reach that conclusion doesn't mean a computer could.

So they're able to infer where I work (the address), my type of job (based on search, stackoverflow is my first result in 50% of my searches) and where I parked the car- but they can't infer my role or the company's business?

Obviously not, otherwise they'd be able to tailor a segment directly to you and sell your eyeballs for way more money than 'mere' retargeting.

But you have to ask, why would Google go to all the trouble of being your personal experience curator? Google is not in that business. They are interested in showing you the highest value ad that they can, and obviously that's nail polish.

Because Google gets money from advertisers. Advertisers in turn pay Google because it makes them sell more. So the more Google is able to provide me with ads that make me buy something, the more the advertisers will pay Google. Taking the money of the advertisers and not boosting their sales doesn't pay on the long run.

You contradict yourself.

> Advertisers in turn pay Google because it makes them sell more.

> Taking the money of the advertisers and not boosting their sales doesn't pay on the long run.

Which is it, advertising on Google boosts sales or no?

> You contradict yourself.

No. Please try to understand what I wrote.

No. What you wrote does not make sense.

How is Google figuring out which Wikipedia pages you visit? I just checked with Firefox's Network Monitor and the only connections are to Wikipedia or Wikimedia domains.

Because most of the pages I visit on Wikipedia are Google search results. Internal Wikipedia links are not tracked by Google, but those are somehow less important (they're related to the first page I visited anyway).

Udik and his responders are talking past each other a bit. Udik is talking about Google + their advertisers as if they were a single entity, and it is dumb for "them" to advertise stuff he's not interested in.

However Google and their advertisers are not a single entity at all. It is efficient/smart for the advertisers to do retargeting (although a cap and funnel adjustments would make it even smarter) and it is efficient/smart for Google to take the money and run the ads.

Udik, ultimately and in the long run I agree with your point; smarter advertising infrastructure will uncover the lost value that you are pointing out. However, given the current structure of the advertising market and state of existing tools, neither Google nor the advertisers immediately stand to gain by changing their approach to people like you. We're in a local minimum/equilibrium.

My biggest gripe is when I explicitly search for a product instead of using their URL and the first ad is the same as the first result. If I go to Google and type in "TeamViewer" the first ad is for TeamViewer, the first result is for TeamViewer... I hope they're not paying for that.

They have to. If they didn't then Teamviewer's competitors could purchase that placement and the next time you search for TV you'll get Citrix or someone else. It might not lead you to click on it but now you'll know you have 2 choices.

> That's silly and a big waste of money.

Most sites that do retargeting do it based on number and stats. They know that if a rando that visits their site (but doesn't buy anything, doesn't sign up for a new letter, doesn't click on contact me) will convert at X percentage if they show him an ad.

So they do the math to figure out what that looks like after impression 2, 3, 4, 5, etc... somewhere there's a point at which the cumulative conversion percentage doesn't go up or goes up so slowly to not justify the cost and they stop.

So what seams like a big waste of money to you, seams like a profitable business decision based in fact to them.

Certainly this seems like the way it should work. In order for it to actually work like this it would require the marketers to actually understand retargeting rather than believing the hype spread by ad networks, and it would require the inventory to be delivered through systems which could provide adequate tracking.

Click-though is often not the expected success criteria for these ads though, instead they depend on overwhelming you with reminders until you crack. This has two consequences, the first is that successful campaigns are absolutely designed to annoy you and not give up at a point which is decent, and secondly the tracking becomes extremely weak, with high rates of inferred success where the retargeting ads may not have had a positive influence at all.

I run remarketing campaigns. One thing they can do is exclude people who visit job related pages. Its small but saves a few bucks (and annoyed interviewees).

it's not based on the search, it's based on the your visit to their web site.