Hacker News new | comments | show | ask | jobs | submit login
Fastest Way to Lose $50 - Our Failed Admob Experiment (fotobookapp.com)
92 points by iamclovin 2269 days ago | hide | past | web | 57 comments | favorite

$50 is not really a marketing budget. You'll get more downloads from this being on Hacker News... oh, wait... clever.

You paid WAY too much.

Let's look at theoretical numbers. Say you have a mediocre conversion rate of 1%. At 30c per click it will cost you $30 to make a $2 sale. Thus, the most you could have expected from this venture is like $3, but even worse, you managed to not figure on Apple taking 30%, so your'e back to $2 in potential money on a $50 spend.


Next time bid a penny and see what happens. At a penny you could profitably sell your app at a 1% conversion rate. That's $1 cost to make $1.40. Not great money, but it's at least got potential and if you got enough volume of sales you could become a top app and make some sick money.

Also, never listen to Ad Networks on what to bid. Bid based on YOUR profitability, not THEIRS. If you can't afford more than a penny a click, then fine, move on.

In retrospect, yeah, I feel like an idiot for bidding the suggested bid. I've run pretty substantial campaigns on both facebook and google adwords though, and always had good experience bidding somewhere within 20% of the bid.

I also maybe was slightly duped by this statement after setting moderately restricted targeting:

"You have selected market area or demographic targeting options. As a result, you will receive significantly fewer impressions."

But thanks, and for anyone reading this, don't advertise paid apps on Admob, and don't bid anything higher than $0.10 if you have a small campaign.

PPC usually does not make sense for smaller per sales pricing. The numbers make even less sense for Amazon affiliate marketing.

Perhaps lead generation instead?

Don't try to initially sell your app, instead, obtain the contact info for your potential customer, and you can sell him/her many different things, including apps, but also books and other products/services.

Also, you will be able to track these customers and know which are most likely to respond to your multiple offerings.

I don't think this reflects poorly on Admob at all - more that paying 30c per click lead for a $2 app doesn't work, and $50 isn't a sample size.

even at a 5% conversion rate from click to purchase (highly unlikely - for this type of app you are talking about less than 1%) your customer cost of acquisition would be $6. do the math.

with these apps you have to be looking at review sites, social media, blogs (like you have now), link exchange with other apps, making it free and putting ads in it, etc. the hit rate is very very low.

Agreed, in retrospect the CPC bid was stupidly high. They suggested ~$0.28 after we set our targeting...I guess I trusted google since their Adwords suggestions are generally pretty accurate.

We also ran a $50 campaign previously though I didn't include it in this post. We didn't even realize we had run the campaign-- 1200+ clicks at a $0.04 CPC in minutes.

We thought our payment had never processed, but when we checked our bank account it had-- it's just that Admob spent our budget so quickly we didn't realize it.

I guess I trusted google since their Adwords suggestions are generally pretty accurate.

All suggestions on the AdWords interface are BIG: Best Interests of Google.

Totally agree with nik here.

Mobile ads are not the means to acquire new customers. Relationships with review sites, blogs, and other media outlets is your best bet. A good review can get you some pretty good uptake. We got reviewed in several Android blogs and reached 18k downloads in a little under a week (our app was free).

I haven't experimented much with mobile ads primarily because they aren't effective and won't ever be effective. At least, that's what the studies say :).

"At least that's what the studies say"

Link?! My preparation for Admob was admittedly limited to one or two google searches and a scan of a few blog posts, but I would love to read more if you would link to said studies.

Okay...now I am a tool and can't find that study...but here is my personal account of reviewers helping downloads for my old application.

We got written up in Thrillist boston, Phandroid, Androinica, Droid-Life, TalkAndroid, and Gizmodo. From all of these review sites we reached roughly ~18k downloads in a week (our product is a fantasy football app).

While the majority of those downloads came from Gizmodo, we definitely would not have been able to get there without being written up on all the other blogs.

From those 18k downloads, we averaged about 11k active through the first 3 weeks of the application and it steadily declined after that (assuming that they were bored of their fantasy league or moved on to another app).

The thought behind the reviewers is if its a source that is trustworthy or has built a following, their post/review is more meaningful than an ad. These reviewers are an AMAZING source of word of mouth for app developers while an ad is simply a pay to click.

If you want to see a chart of our downloads over time, I'd be more than willing to share to show you the value of these relationships as opposed to a one and done ad campaign :).

I think you're missing his main point, which was he set it to spend "evenly throughout the day" and it was all gone in one minute.

I'm not sure how that makes a differencee if you're paying CPC. x clicks is x clicks, regardless of when they are clicked.

If you're paying CPM, then sure, care about how spread out your adverts are.

Emotionally shocking, yes, but what is really expected? Why _should_ the ads be spread "evenly throughout the day"? esp. for such a pittance? The spread has no advantage. Google is handling a deluge of data, and he wants a controlled long-term distribution of a few miniscule tidbits. Yes, expecting Google to behave an obvious way makes emotional sense; methinks it's all just happening on such a vast scale he/we don't grok that blowing $50 on ads in seconds does, in fact, make sense.

I just thought they should be spread throughout the day because I selected the little radio button that read:

"Spread ads evenly throughout the day"

Mobile ads are largely a money drain. Most clicks are accidental and will not lead to a sale. I'd only recommend using AdMob if you are promoting a free app and are doing a burst campaign to break into the Top 25 chart of a particular category.

Secondly, a $50 budget is peanuts and paying $0.25/click for relatively un-targeted admob clicks is crazy. We normally pay $0.05/click for such crappy traffic.

We've spent 1000's on mobile ads and it's only an effective strategy for pushing your downloads a bit higher so that you have a better chance of hitting a 'gold spot' in the charts. E.g. when you are 26th and you want to get into top 25.

I agree. The ~$0.25 CPC bid was suggested, we actually only decided to go with the suggested bid because they had warned us that demographic targeting would limit exposure.

We actually ran a similar campaign a few days previous where we had over 1,000 clicks at a .04 CPC and didn't see any increase in downloads.

From a 1000 clicks i wouldn't expect more than 50 downloads of a FREE app, for a paid app, you'd be lucky to get even 2 or 3 downloads.

Fair point. I think the lesson learned was that the targeting that we thought we were employing was not nearly as effective as we had anticipated (25-44, limited geographic area, iphone 3GS, iPhone4, iPad2).

We also were just amazed at the clicks registered in just seconds. The stat reports were already in and email alerts received in less than 2 minutes time. (Daily Budget and "Spread throughout the day" notwithstanding)

btw - did you use AdMob's download tracking?

I ran a $50 Admob campaign over the weekend for a new FREE app we have on the Android Market. It did exactly what I needed it to, it got me about a hundred downloads. Today, that app has a little over 400 active users.

The ball has started rolling, and it only took $50.

THAT BEING SAID, the first time I ran an Admob campaign, I had just about the same experience as this writer.

Admob works great for getting lite-version, free, download apps a little traction. I will never again run it for a paid app.

IN-FACT, in the Android Dev Console, you'll notice that only your free apps have the "Advertise this App" link next to them. I assume that Google probably is very aware that Admob campaigns can work great for free apps, but it's a different ball game when it comes to paid apps.

Lastly, I'd like to add that I took the time to create a graphical banner for my free app ad that made a BIG call to action - Download my new puzzle game for free.

Thanks for sharing. Your conclusion for the majority of app developers is spot on-- for small-medium market paid apps, Admob isn't worth the money.

We're probably much better off spending it all on our Free version with limits to prove the value to free users and hope they upgrade.

In retrospect (and maybe this should be a follow-up experiment), we should have advertised ONLY our free, limited version and focused on getting the cheapest CPC possible.

$50 at a $0.02 CPC would yield 2500 clicks. For a free app, if we saw a 10% download (a very healthy assumption I know) that would be 250 downloads of the free app.

Maybe after proving the value of the app with limits, we get approximately 10% to convert to paying of the 250 downloads.

That's 25 paid, or ~50 pre apple tax. Still losing money with very healthy assumptions, but I think we'll try this as well...even after the negative experiment and will post our results.

I just want to say thanks for posting your experiences. Everyone here is ragging on you for "too small sample size" and "u did it wrong" but I appreciate that you came forward and shared.

For the people who scoff at a $50 ad budget, that just tells me they've never done this with their own money. It's easy to brag about campaign size when you're spending out of someone else's wallet.

If you want guaranteed results, use a cost-per-install campaign. I work at an ad network that specializes in these kinds of campaigns.[1] You'll have to pay a lot per-install, considering how poor conversions are going to be. Our network automatically calculates your eCPM on the fly and will rank campaigns according to that—and so too it works with most competitors' network.

To be honest, with your tiny budget and a two dollar app, there's no effective way to spend your money with a positive return on investment.

[1]: http://chartboost.com

Doesn't Apple's policy not allow Pay-per-install Ads or something? I remember Tapjoy ran into issues because of that.

They don't allow incentivized downloads. (Ie. we'll give you in-game credits if you download the app that this ad advertises.)

The only ad platform I have ever had any success with is Reddit's cheap $20 self serve ads. I haven't tried any Facebook ads though so there's that giant I have yet to try out. Reddit's great though mainly for the feedback and people who are actually interested in what you're selling and willing to critique/x post your ad.

Interesting. We did think of Reddit but maybe thought that an iPhone app which is a Facebook utility is probably not a Redditor's cup of tea.

I experimented with AdMob and JumpTap over a few months with budgets ranging between low tens to low thousands for each campaign and even for an app that was $5, it basically didn't work.

I didn't get good results with Facebook either, but maybe due to the nature of your app, it'd be better.

Ditto for Adwords.

I only dabbled slightly with Reddit and it seemed that it would show some promise if I invested more time in it.

You might also consider paid downloads promotions. Something like http://tapzilla.com/ (YC S10). The co-founders Edmond and Francis are both helpful and responsive when I run a campaign. It wasn't the right avenue for my app, but Fotobook sounds like a better fit, as a utility and at lower price point. I'm not sure if Apple's recent foray against paid downloads affects them.

Thanks, we'll check out TapZilla.

Amazed that people will spend so much time and energy trying to save $1 but won't spend time trying to make $1...

Would love to see a blog post about your experience with reddit traffic data / demographics etc...

I'm assuming you can target subreddits with ads?

To help capitalize on all the HN traffic, you might want to include a more prominent link to fotobookapp.com (probably in the first sentence when you introduce the app). It took me a while to see the link at the end of the article.

Also, the "App Store" link in your posterous profile doesn't actually link anywhere.

Thanks Michael - fixed the broken link.

How much does news.ycombinator.com charge per click ? What is the conversion ratio of news.ycombinator.com users who clicked on the link ? ;-)

Yeah...maybe we'll make our money back.

But this comment stream has taught me a alot, we want to run a considerably larger experiment with our free application and will report back.

FWIW, $50 is a very small amount to spend on advertising like this. So small that I don't think you can draw too many conclusions from it.

That's very fair, we're just submitting this as a benchmark.

We also (mistakenly) also spent $50 on Admob a few days before this. We didn't even realize we had spent it, that's how fast it happened.

I just can't conceptualize technically how these stats could be real.

A year ago, Google was doling out results at a rate of 34,000 searches per SECOND. In your case, that's filtering and spreading your ad across some 3.4 million searches. Not unreasonable to assume your demographic restrictions fit about 1% of those searches; the limits sound narrow, yet seem candidates for the largest group of Google power users (whom I would independently assume are dominated by 25-45yo upper-economic-half US state residents owning camera-laden iOS devices).

You're trying to take a drink from Niagara Falls and wondering why you're soaked. (Been there, done that. Memorable, but not productive.)

I guess that is normal. 150 clicks and 0 downloads for 1.99$ app.

I think around 50+ clicks were accidental, and immediately back key was pressed before loading the content.

And with other 100, i guess your assumption was wrong that there is big enough market for this.

But something else is interesting here, you really can make money on ads from free apps --

In the past, Facebook Canvas applications and Website applications have received C&Ds for the "save photo" functionality.

See: http://www.vincentcheung.ca/blog/facedown/cease-and-desist/

I do not know how the author's application retrieves Facebook photos, but I believe that Facedown (which was a desktop application, not a web app) was shut down because not because it was saving photo albums, but because it required users to input their Facebook credentials into it and tried to download photos by scraping album information (by pretending to be a web browser).

I have operated a reasonably popular Facebook application that has similar photo saving functionality (using their API), and not have not run into any issues. It has transformed from a desktop application using the REST API to a Graph API using web application at http://socialphotodownload.com . Unlike Facedown, I try to follow Facebook's Platform Policies.


It seems ridiculous to me that Admob ignored the $15/day budget and spent $50 in two minutes.

Yeah, see some of my other comments. I emailed Admob about this, and their response was:

"Please note that budgets under $200 per day can still be spent in a matter of minutes."

How does it make a difference at all? Unless what you are promoting can't take the traffic, it's irrelevant. When you're paying CPC you have no reason to care when the clicks come in, 10 clicks a day for three days isn't any worse or better than 30 clicks in five minutes.

> How does it make a difference at all?

It's what was promised. If they can't limit a campaign to a certain dollar amount per day, they shouldn't offer it.

It makes a difference because it's an explicit setting. $15/day. And who knows how that scales? What if going through $500 only takes 45 min?

As long as it doesn't spend more than the customer wanted to spend, it's irrelevant.

Let's say I want to spend. If I have $500 to spend, and I want to pay CPC, how is getting $500 worth of clicks in 45 minutes worse than getting $500 worth of clicks over three days?

I think it is relevant because it's wrong. Unless it says somewhere that it's a suggestion or it probably won't work like that.

If you tell AdMob to not spend more than $15 a day and then it immediately spends triple that amount I wouldn't say that's giving you a fair shake on what they promised.

AdMob sucks. IMO they do this on purpose to discourage small publishers. I've also heard, from a reliable source, that they also send you worse traffic until your account is more "established", and if your early CTR sucks, you may never get the bump. I've heard of people having to start new accounts after their first few campaigns do poorly.

I suggest you try InMobi and AdModa. (Not affiliated with either except that I've been happy with the traffic I bought there)

Because you might have a $500 testing budget and want to test how your campaign does over the course of the week. You lose out on valuable data even if your $500 in a few minutes converts well.

I liked the first comment (in the post) by Andrew Maine. Basically, it's simple math. If you expect a 5% conversion on a $1.99 app (which will earn you $1.39 after Apple's cut), the most you can spend for a viable business model is 6 or 7 cents per click.

Less if you want to make an actual profit on the app and not just get more users. But then, if you want more users, maybe the app should be free?

But once you make a conversion, do you have any way to subsequently contact that same user to make them aware of your other app/service/product offerings?

Also, if the app is recurring billing, the formula changes completely.

I know I'm not the only one who feels really sorry for the fellow. I'm imagining him sitting by his laptop and scrupulously making sure his target demographics are correct for his ad campaign, and this nonsense happens. Google should make it clear in the FAQ that your campaign will be short lived if you ad funds under a certain amount.

maybe the stats are fake. i think thet want you to buy 5k minimum to get some start of a result

We actually got a clarification from Admob that I should include in the blog post:

"[P]lease note that budgets under $200 per day can still be spent in a matter of minutes."

I was just frustrated that they didn't include this anywhere in their campaign or ad group setup or their FAQ. I had to write them to get this answer.

And then, why even allow for daily budgets below 200?

You can try things like getjar.com which offer a cost per download model. Or list your app on my directory @ mjelly.com!

Applications are open for YC Winter 2018

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact