I Spent $6M on Google Ads Last Year
nicklafferty.com
nicklafferty.com
If I try to optimize for download volume, that's all I get - bots after bots. Can rack up thousands of downloads for <10 cents per download without any impact on how many users actually use or rate the app.
I found that getting a good ROI required setting higher price targets and optimizing for in-app actions. That's really hard to do well and not very reliable, but it helped.
How does a bot know whether you've optimized for download volume or in-app actions? I don't understand how a bot would know to download the app to fulfill your requirement?
Furthermore, what is the incentive to write and operate these bots? Genuinely curious. The only party with an incentive to do this would be Google, no?
So if bots are more likely to click download, and bots have similarities, bots see more ads.
Google has a strong incentive to let them.
If you've ever taken a look at the number of conversions fb ads provides vs google ads, you'd know this.
Edit: that being said, this does not seem to hold for Facebook, and I don't have a good reason why.
Google's advertising business is built on trust. The moment there are a few scandals involving fake/robot ad clicks, big advertisers will remove their ad campaigns.
So while it is risky, not sure how they could be caught. I don’t think they are running click fraud bots, but it’s hard to quantify how hard they work to stop them.
A friend of mine estimated that 1/3 of ad spend is click fraud but it’s just in the wash because it’s hard to stop.
I'm assuming you use display for this pricing? If so build a white list of acceptable sites or avoid display. Running open network is always waste of cash in y experience, especially mobile devices.
I also just released my app on the Play Store a couple of weeks ago, and had exactly the same thing within the first 48 hours - a bunch of suspiciously similar Gmail accounts, none of which had actually used the app (I know this because no backing accounts were made in the database).
The only conclusion I can make is that there's huge amounts of fraud, but the average advertiser either doesn't know about it, or their advertising ROI is high enough that they can simply eat fraudulent clicks as the cost of doing business.
I've said it before, though - gut feeling is that internet advertising is long overdue for a reckoning.
The Downloads folder is a scourge that never should have been allowed to exist.
Where should something be after I request it be moved to my machine? On the default view! (Aka Desktop)
Why? Because I literally just interacted with it!
Microsoft seemed to think shoving everything into a junk drawer was a better solution.
Let's say you see 75% bot traffic. You make $10 per actual user. If youre spending $2 per action with 4 actions, you burn $6(75%) on bots, $2(25%) on a real user($8 total), and get $10 - netting $2. So you're making money, but wasting more.
So, let's say you set your targeting to completely exclude the bots. Great, you spent $2 and made $10, netting $8. The problem though is you ruled out 90% of your traffic - you have false positives on whether its a bot so you exclude legit users(all of this targeting is fuzzy rule based). You can 10x the traffic with bots, making $20, but you cant 10x the perfect targeting, so youre stuck with $8.
- When you say Adwords for search, you mean google.com and not display right? I see a bunch of what you describe across the display network any would only recommend running white-listed sites if doing the latter as there is much encouragement of fat finger and kids who press anything etc here.
- Are you allowing search partners? Id going google search I would strongly recommend sticking to google only as there is often more dodgy around partners but this is auto included in setup.
- What kind of KW and variance: Again assuming search, who big are your KW lists and negatives? Apologies if asking the obvious but is this consistent via KW?
- Geographic restriction: Have you broken the effect down geographically? This can lead to huge variance, and to a letter extent times your running.
- Are the ads pushy? Google and some people will put a bunch of effort towards you getting higher click through but fight that. You need to stay withing bands but too many ads are for an off the cuff example; 'Amazing game that will blow you away" gets clicks whereas you'll perform better by setting expectation with the boring 'Download and install this game' as your headline type thing... naturally if bots it wont help.
Game install can be tricky, from what I've seen it's competitive and low margin which is a challenging start point.
It's the same big giant list of keywords I used to use, something like 200 long, which I optimized a lot at the time. Not a lot of negatives, though.
USA and UK only this go around.
I don't think they are pushy. Yes, I noticed in the past that honest, informative ads seem to work the best, especially since I used to do CPC, so no reason to trick people into clicking who really aren't interested.
I'm sort of in a weird position where some of my metrics seem really good, monthly ARPU of $9.86 and churn 1.5%, I just can't figure out how to get people in the front door any more.
It seems counter-intuitive 'Why would I add more work for people who want to use my product?' - but the truth is you're just creating more work and wasting resources qualifying poor leads (in your case bots).
e.g. each install contains a unique slug, when the installer is opened you use the unique slug to know where it is coming from?
Edit: The installer checks our server to make sure it has the most recent version, and that's where we compare IP's.
Don’t underestimate fat-fingering
Maybe a link to some sort of "app store" (or whatever its called on Windows) would be more convincing?
I believe all of my current software besides games were installed this way.
(More than happy to hear suggestions for easy-to-use sandboxing for desktop software. The whole situation makes me quite uncomfortable.)
(I should probably change my ways.)
You cannot imagine how great AdWords was in 2004. It was much more transparent and easier to use, and the ROI was just incredible. Their success at that time was more than justified. Not so sure about today, though.
Reference: https://support.google.com/analytics/answer/2763052?hl=en
So unless you have additional identifiers, an IP address by itself isn’t enough.
Article 2(a) of Directive 95/46/EC of the European Parliament and of the Council of 24 October 1995 on the protection of individuals with regard to the processing of personal data and on the free movement of such data must be interpreted as meaning that a dynamic IP address registered by an online media services provider when a person accesses a website that the provider makes accessible to the public constitutes personal data within the meaning of that provision, in relation to that provider, where the latter has the legal means which enable it to identify the data subject with additional data which the internet service provider has about that person.
Additionally, from a cursory read, it appears like they are only talking about dynamic IP addresses anyway, but you, as a service operator, have no way of knowing if an IP is dynamic or static.
I never really understood how Google could ever guarantee that ad clicks are real and not fake, as any sufficiently sophisticated robot will have click patterns that are indistinguishable from humans. I bet designing such sophisticated robots is an industry on its own.
For example, could you build a page that bots could navigate to, but would be really obscure for a person? Would that generate a substantial amount of clicks? Sure, the bots could adapt (i.e. not click on obscure pages), but that would take time to learn, and until then, they would screw up and fall into the honeypot.
If that has ever happened in the history of Google, you'd think someone would have publicized it.
Got a bunch of phone numbers that wanted our service. Only 1 person was interested, but didn't sign up.
So you end up boosting or wasting money on ads. In my (limited) experience, ads weren't effective for us as they were too local and the audience was limited. Boosts were effective but expensive.
The funny thing was the our highest conversion paid event advertising was the old fashioned newspaper.
There is no mention of actual ROI, conversions, traffic, and so forth.
Overall, I'm pretty skeptical of the true return that marketing campaigns can generate. When you calculate the full cost not just of the marketing but you include margins and the full lifetime value of the customer then you get a real sense.
The thing that stood out to me is the first sentence, that the person had an unlimited budget. With that in mind, basically makes me believe they weren't that focused on the true profitability of the spend.
No one has a $3m Google Ads budget for something that doesn't work already. I've run campaigns for brands that lead the market and for new businesses that have no brand recognition at all--the difference is night and day. Should founders run to Gmail ads because of this? Probably not.
Assuming the campaigns are still profitable, doubling the budget is great. I hope they really did well. But what does this really tell us? What can a startup or small business founder learn from this? Not much.
Blake, Nosko & Tadelis, 2015. "Consumer Heterogeneity and Paid Search Effectiveness: A Large‐Scale Field Experiment," Econometrica.
https://econpapers.repec.org/paper/nbrnberwo/20171.htm
I'm sure the landscape has changed since then in terms of there being more accountability (better tools to measure things like ROAI and mgmt demanding stronger evidence of marketing efficacy).
The agency that runs the programs the the only people that win - after they pocket their 5+% commission.
Source: past roles where I saw a lot of people who do not know wtf they are doing with ads (yet they themselves are making 6 figures)
Adding even more nuance, that really only describes successful companies. Failing companies _think_ they are profitable, until they quickly aren't. Your experience is very similar to my own. You're 100% right.
The ultimate goal of the marketer is to use the most effective, scalable channels. The optimization isn't against a fixed goal (CPA), but against the entire universe of ways to spend each marketing dollar.
- This was my previous job, and I don't have the data anymore :) - I don't want to reveal numbers that could harm that company or inform their competitors.
I will say I was 100% focused on profitability. I'd be a pretty shitty marketer if I doubled spend without focusing on profit.
The past few years, Google has been adding features that make absolutely no economic sense (assuming rational firms), but assuming that most of their clients are unsophisticated and then taking advantage of them has no doubt earned some Google Ads PMs huge bonuses.
My first thought was not "I was in charge of spending my employer's money on ads", it was "I was spending my money on ads".
I wish there was a clean, sharp division between people trying to sell you stuff on the open internet and entities that you are logged in to and already doing business with, such that the latter were guaranteed to be accessible even with maximum blocking.
Exploiting customers is something that needs to be regulated more.
I can only count a handful of sites that break with conventional blockers (uMatrix might be a different story) and those sites I can just avoid doing business with (only one I can remember recently is Patreon where the Facebook/Google ad scripts cause the entire page to be unusable).
I've installed them for a couple of friends and family who complained about ads and slow websites, but hadn't known about blockers.
Whenever I turn off my adblockers and visit common sites, it feels like I've been wearing hearing protection in a loud place and I'm now taking them off.
I think it depends on the Inbox type you use as well (e.g. Priority, Default etc...)
One thing he doesn't mention is the landing page experience. It affects both your quality score which can have big impacts on CPA as well as your conversion rate.
Not sure why people have issue with "creative" as a noun, this has been the standard in advertising for decades. Hell, it's in the dictionary: https://www.merriam-webster.com/dictionary/creative
Don't need to crap on Quantic and the like as much though, they do have some things that Google doesn't.
Some people are commenting on him not giving specific advice for making the creative. As some have pointed out, a whole article could be written about that, or more realistically, entire libraries of books. In short he is talking about ad copywriting, it's a whole discipline that takes years of experience and learning to get good at.
Most people aren't in advertising. It is a surprising use of a common word. I still twitch when I hear 'copy' used in the content sense.
Probably the same problem some people have with the term "Human Resources". If you can be reduced to a "Resource" then you're basically just meat (or it can feel that way).
When you use the term "creative" as a noun that way it sounds like a brain in a jar, and if the brain misbehaves, it can be replaced.
This bidding mode means you're taking less risk per auction, as you don't pay for impressions or clicks. You pay (sometimes 10x more) to reduce that risk, though, and your reach (how many people see your ad) can be adversely impacted if the Google Brain estimates your click through and conversion rate unfavorably to your auction competitors.
Anyways, action can be whatever you define. It could be a purchase, someone signing up for a mailing list, or someone sending a purchase order, or whatever you can programmatically define.
A click is just opening the landing page from the ad.
The design of the budget system is in Google's favour, not yours, and if you don't understand that, you'll end up paying much more than you need to.
If you think setting the budget to half is a way to halve your ad spend and get half the number of clicks, you're right, but you're also naïve.
If Google thinks your current rate of spend is going to exceed the daily budget, they'll effectively deactivate your ad for random impressions, leading to less spend.
but that isn't the best strategy. The best strategy is to lower your bids to win fewer auctions if you want to spend less. That way, you can spend half as much, but get more ROI, since the Google AI will be only showing your ad to those most likely to be interested in it.
Manual bidding is always preferable with fewer data points (low conversions, low clicks, no previous campaigns). Auto bidding is preferable as your GA accounts grow in conversions and the algorithms start to create a better profile of your customers.
Bids are what you are thinking of, and I'm afraid this is the nature of an auction...
Remember it isn't highest bid wins... It's highest bid multiplied by a secret quality and relevance factor that wins.
If your ad is more relevant and useful to the user, it can win with even a very low bid.
IIRC, some of its major channels do it when the campaign's going to exceed its budget. Or you can simply use a maximize clicks/conversions bidding strategy.
Most businesses have some marginal cost of providing service to a user, and are therefore happy to pay any amount up to that cost to get a user. So in the Google Ad dashboard, that would be "Target CPA", but instead of setting an average cost per acquisition, it would set the maximum cost per acquisition.
Whats the difference you ask? Some users will come via a 'cheap' route - for example directly searching for your company name, and clicking the top ad, costing you very little because your quality score is very high, and then converting at a high rate.
Other users will come via an expensive route - searching for a tangentially related keyword, unlikley to click through, and also unlikely to convert when they have clicked through.
You only want the guy from the expensive route if he, and all the others like him, cost less than your companies marginal cost of providing service. You don't want to include the guy who directly searched for your company name in the average when making that decision.
Why doesn't Google offer this? It would significantly cut down on overspend in the ad world (ie. spending more to get some customers than they are worth), and in turn significantly hurt their revenue and the internet as a whole.
3rd parties that use Googles RTB API can offer this though, although they are at a disadvantage of having a lot less private user data for targeting...
Just disable broad matching
https://www.statista.com/statistics/266249/advertising-reven...
Personally I don't like these things because I know what I want before I buy it and I research almost everything I buy, and I probably have some anti-consumer bent deep down in my belly that stigmatizes shopping for the sake of shopping.
But not everyone is like this. I've heard people say they like ads before, and AFAIK it's usually clothing ads.
Now that facebook has embedded ads that get past OOTB ad-blocker and are super targeted to me, I've actually clicked on a couple ads that have intrigued me, knowing I'd not buy it but interested in what their pitch was. Most of those were kickstarter projects or some kind of "adventure gear".
Sometimes this will turn into anger toward a brand even, like when I'm visiting family over the holidays and I momentarily remember just how bad cable television ads are. I just start to legitimately hate a company because of their ads. I wonder how many companies calculate value lost from an ad because of annoyed people.
The difference between ads on HN and ads elsewhere is relevance. You self select for useful ads by going to HN. Because I regularly search for things that I actually need on Google and youtube, I end up clicking on lots of ads on those platforms too.
Impressive
The author is spending $6M a year on ads and is generating substantially more revenue than that, they claim. Why are they bothering with spare-change Amazon affiliate links? If they can get revenue like that, they should be making tons of money with their day job. What's wrong with this picture?
These days many top performing video ads are the ones that are made to look amateur/cheap. And it's not just the guru course ads. The native vlogger-style ad is often more engaging on a platform filled with vloggers. High production value "professional" ads have traditionally been created to lend credibility & authority to the brand. But if their audience now trusts "influencers" more than big brands, what's the harm in testing a TON of simple video creatives with clever scripts.
I've even seen this style now in offline commercials.
If you don't like "native vlogger-style ads" don't watch those videos.
> If you don't like "native vlogger-style ads" don't watch those videos
No. If the content creator wants to give access only to the people who can provide them income, then they should make this transaction direct and transparent. It is not on the consumer's job to ensure that the producer is paid.
Is that accurate? If you follow a link to watch a video, is that not asking to watch that video?
When a site serves ads and tracking from its own first-party domain and I can reasonably determine this, I do not block it and often unblock it if it defaults to blocked.
When a site farms out adverts to third parties to share my browsing habits across the web, I default to blocking everything.
If there's a site that decides that it won't show me content because I'm using an adblocker, then I move on. They don't want my eyes, and I don't want their practices - no reason to stick around.
"Creative" in the context of an ad means the image, video, text, etc. Basically, the ad content that the user will see.
The target market there is different though. Think instead of a single company spending $6m on their ads, more like an agency spending $60m++ on 10++ different companies' ads.
Major props to this guy for doing so well in house though.
Ultimately, this needs to be worth the price though when their fees are measured in % of media.
Hiring creatives to make your creatives.
Having people all over the world shown an ad for a specifically localized shop ? No thanks !
For more online businesses, the results are indeed quite decent !
The third party services aren't much help. Google has a lot of great in-house support to ramp up your Adspend once you get noticed. Not sure what the target is, but we seem to get a new rep every few months who has new suggestions.
I have no experience with Youtube, but that is highly interesting.
If Adwords isn't working for you, there is a ton to dive into. Could be your website, could be your product. If you have a competitor obviously succeeding in the space with heavy Adwords spend the fault is probably on your end (unless they're just pissing away money, but how long does that go on for?).
Very different from taking something already successful but mismanaged - this can be a lot easier to ramp off if they're making big mistakes (like not actually mobile optimized, not CPA bidding, not actually actively optimizing the website's funnel/landing page).
I somewhat disagree with the author that the creative is everything - unless you already have all the other aspects locked on and creative is the only place for you to improve
So what is the best way to iterate on new creatives without re-entering the learning phase?
The ad is the medium, the creative is the message.
One designs a message: "people who drink $cola are fun" and then the creative to convey that message - such as an image showing someone smiling, nestling the product in their hands.
It's a shorthand term (think DSL) a bit like "code" in the world of developers, etc..
Are there good online courses or blogs that you recommend?
If you want to learn Google Ads I'd recommend getting certified for Search, Display, and Shopping. If you want to learn Facebook and Instagram ads then that's a whole different animal.
to give you one back, my cringe dev jargon of the moment is PERFORMANT
I used to be in advertising and “creative” referred to the role of creative direction. “The creative” would be a person. Had never seen it used as a synonym for content.
He is spot on about the landing page but there is one more thing I learned that played a key role and that is something called the keyword smell. Basically what it means is your landing page must match the keyword you're targeting. If you're bidding for "home windows for the elderly" but the landing page headline says "home windows" you will lose the lead. Plus google will also never optimize your ad for clicks. Even if they have to optin and go to another page just don't lose that smell. It can work wonders for your campaign.
A bit like is exercise more important, or diet? Well if you are eating only lard and running every day, then diet is important for you, and if you are eating really healthily but sitting in a chair 14 hours a day, then doing more exercise is the impactful thing.
Plus, users who are actively searching for something are more primed to respond to an ad that appears on the results page, as opposed to an ad that pops up in the middle of a news article they were reading. Users for the most part don't seem to like stuff with animations, or heavy graphics in their ads as they're distracting.
I have noticed more interesting ads on newer mediums (media?) like Instagram stories or Snapchat, which are image and video-centric. The targeting so far seems quite good, as I usually see ads for online courses, creative stuff, design tools.
One notable one only ran once a day, and simply used data from your account to adjust other things in your account. And it used a crude algorithm to do so.
You literally could do better with a shell script running hourly after reading a couple of papers on the multi-armed bandit problem.
If you have another solution apart from google ads, I’m all ears. Seriously.
the solution is to inform your audience how they should forcible disable ads and to switch to a search engine that can turn off ads.
Does Gmail have ads? I know it used to, but I can't remember the last time I saw one. Looking at my Gmail inbox now, and I don't see a single ad anywhere on the page. I'm not running any kind of ad blocker.
Disclosure: I work on the team that builds the client libraries for this API.
The other day I spent some hours trying to find reports to show where this friend spent his adwords money and just about every report in his dashboard was blank. He has been running this campaign for about 2 months now - so there should be some data.. I only found one report that showed any data, the rest blank - so then I went to help, then chatted with adwords support - explained the problem, what I was looking for.. told to wait some days for a response cuz covid.. okay. Got an email forwarded to me saying that you can't get that data with 'smart campaigns'... "I would like to inform you that, you can view your search phrases performance (impressions, clicks and status) in the campaigns dashboard, however, cost option is not available for smart campaigns."
So... this is disturbing - and then there is this "You may use the search terms report to see how your ads performed (clicks, impressions, CTR, cost, etc) when triggered by actual searches within the Search Network. "
Yet I have clicked virtually every link in the dashboard and I find nothing of the sort.
I feel like I can put about 2 more hours into this - and then my advice will be to just to cancel adwords it's too opaque.
So should I spend 2 hours learning this API thing, or trying to re-learn adwords to switch off smart campaigns, or a secret how to use reports with campaigns doc or what, I don't know how to get the info needed.
I have seen a few episodes but I feel very weird about taking anything from it other than entertainment. Problem is if you are already subject matter expert, you can distinguish the information from the entertainment but you already know the information so it might be a refresher at best. For example, doctors and other medical professionals watching House, MD or Grey's Anatomy might be able to relate with some of the things when the show has some basis on reality and laugh at things when they are off base. Similarly, people in marketing/advertising business might be able to tell what is what but from the outside, I can't tell the difference between information and entertainment.
This isn't an unusual mode of commentary for HN, but one of my resolutions is to point out when social cynicism has far out-stripped reality.
I guess this is the main reason why Gmail exists.
In media circles, a "creative" might reference an artist, though, not their result of their efforts.
I was being genuine in my first comment. In a time of a pandemic and global depression, I guess you really should just expect the worst from people.
E.g. instead of saying LMGIFY, the response could have said that a creative is the thing that is the audio, text, picture or video part of an advertisement. Or as M-W puts it, "creative activity or the material produced by it especially in advertising".
Your question could have been improved by saying what exactly you didn't understand from the first few Google results, especially since there's a clear definition of how creative is used as a noun in the first couple links.
A good question includes the context of what you've already considered.
https://www.merriam-webster.com/words-at-play/can-creative-b...
I'm glad that a stranger being snarky to you online is the worst thing you've experienced in the current situation.
I hate the modern internet. Platforms are holding developers hostage. Because if you don’t pay for ads for YOUR PRODUCT then they’ll show your competitor’s ads.
It’s anti-consumer hostile. If I search for a thing I want to see that thing. Not a different thing that has a bigger as budget.
What he is saying is that nailing the actual ad trumps everything else (targeting, keywords, platform, etc) when it comes to being effective in advertisement.
Source: https://www.merriam-webster.com/words-at-play/can-creative-b...
It's a gussied up way of suggesting this is more highbrow than 'ad' which tends to have a pejorative connotation.
"How did the ad we put up last month do" sort of conversations
The heirarchy of terms gets really confusing here
"Creative" is more succinct, but I don't think it's clarifying anything.
Creative = Story