Show HN: Ad Revenue Prediction Tool for Side Projects
marginhound.com
marginhound.com
#1. The tool estimates 1.5 million a month in revenue. Actual earnings are 6,000.
#2. The tool estimates 6,600 per month in revenue. Actual earnings are 400.
I get those numbers while using conservative prediction options. With the first site, the estimate is 250x higher than actual revenues. It's hard to even call it an estimate at that point.
I'm confident my sites are well optimized. I run header bidding with most of the major ad networks (Google, AppNexus, AOL, etc). I've done countless A/B tests. I've sold direct campaigns to Fortune 500 companies. I frequently Skype with other well known people in my industry (they're pulling in millions annually on advertising) and we talk numbers for our different projects.
In short, I just want to caution people these estimates are likely incredibly inaccurate.
The model behind tool was based on a decently large / diverse data set of small websites; it lines up well with public information from sources such as site brokers.
(and of course, I run sites as well and plugged my own data through the tool during the validation process)
What kind of traffic are you getting on those sites?
Probably also worth asking if there are site design or traffic factors that reduce advertising interest? (for example, forums and downloads may have lower intent)
Also, don't forget about ad blocking. Over 50% of my users block advertisements, since they're young and tech savvy. That's a big drop in revenues right from the start.
The first site has about 30 million page views a month from 3 million sessions. The second site has about 1.5 million page views from 250,000 sessions.
Typical CPM rates...
* $0.20 - Ad exchanges for low value regions.
* $1.00 - Ad exchanges for high value regions (US, UK, Australia, Canada, etc).
* $3.00 - Direct sales for MPU spots in high value regions.
* $5-10.00 - Direct sales for background advertisements in high value regions.
This is in the entertainment niche.
Yup, blowing through a bunch of impressions on the same audience will definitely tank CPM's and fill rates. I'm also going to guess you get a lot of repeat visits? (so those sessions are hitting a smaller base of uniques?)
Hmm... thinking I may work up a corrective adjustment.
Core development sample was blogs and small content sites, where the opposite problem exists (finding a way to unbounce the visitors so you go from 1.3 PV/S to 3+).
Ads are psychological malware, even when they aren't serving actual malware.
> why so negative?
Probably has something to do with the advertising industry doing everything it possibly can to infiltrate every facet of our lives and ruin everything that was ever good about the internet and the web.
I disagree at multiple levels.
While ethically attractive, most people don't pay. This has been proven multiple times across multiple mediums. Look at free-to-play gaming, for example. Worse, go see the badgering and feature locking which goes on as they attempt to extort you into forking over some cash. Is that really the web you want to live in?
From a creator perspective, this sucks even harder. Now I have to manage a bunch of paying customers, many of whom want "support" since I accepted their cash. Some of them even will attempt to bother me in the offline world over any disagreements between their expectations and my own. You would be disturbed by some of the emails I get from patrons of "free sites" - god forbid I should take money from these people and be morally obliged to serve them. My favorite part about being an ad supported publisher is having the moral freedom to tell a jerk to hit the road. Or call the cops if they attempt to contact me offline. I really don't need another boss in my life....
A good ad optimization process can provide a ton of value to the user experience... because keeping users happy is directly related to making more money. Higher bounce rate is lost revenue. Less pages per visit is lost revenue. Slower loads hit you in wallet too (higher bounce rate, shorter time on screen for the advertisement). Oh - and Big Daddy Google turns off your organic search traffic if you deliver a really lousy user experience. Game over.
Now... most of the ads that annoy customers are actually very low paying placements. Those click-bait sites? That wasn't high paying ad space, for either the click which sent you there or the clicks you did while you browsed. Browser redirects? Also cheap, throwaway traffic.
Most websites would be better if they eliminated their lowest advertising placements. The quality usually sucks.
If it has no value, stop accessing it instead of blocking the ads.
The exposure you are talking about is a non-issue, particularly given every site you contact has that same information (even without Ads).
If you are concerned about that level of privacy, you should be using https://tails.boum.org or a no logs VPN.
Comcast and Cox can afford to set up pay-walls and service packages to create content and fill the void. Little guys like me will probably go find a new hobby like camping or brewing craft beer.
The irony is the worst ads actually get the lowest rates; disruptive formats are highly toxic to a brand marketer (why should I put my name on something that pisses the consumer off)... most high end brand players want to get native ad placements or "editorial looking" ad formats. Surfers have learned to ignore people "yelling louder".
I am hoping IOTA can bring this.
sessions per month: 43000
pageviews per session: 2.3
predicted ad rates: 7.5cpm
actual ad rates: 0.72cpm
actual ctr: 0.27% (280 clicks)
predicted revenue: 742
actual revenue (november): 73.87 (carbon ads)
wonder if it's really possible to 10x ad revenue with a tracking ad network. I suspect this might require much more intrusive ads though.