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?
You don't have to be old -- we still use FTP all the time at my job. Of course, whenever we have a say we use some secure variant, but you don't always have a say (no matter how much you yell, grumble, sermonize about the dangers of insecure tech, etc).
Of course, you probably meant you feel old because "FTP" didn't immediately register to you as "Free To Play". Well, I can't help you there, because by that standard I guess I'm pretty old too.
:-D
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.