Home > Paid Traffic Sources > Mobile

---

How much to spend on a testing campaign to be considered "enough"? (21)

---

Hi guys. I want to know your opinions about testing budget for a single campaign.

I created one campaign promoting UC browser. I paid for the translation, set up everything carefully. I run 30$ to collect data and after seeing the ROI was -70%, I abandonned this campaign. 10 days later, I saw my lander on spytool. someone cloned exactly the same and now sending high volume to it (nearly 10k times seen on Ads Exposed). I was suprised indeed and felt a bit mad at myself.

Someone please tell me, how much should I spend for one single campaign to be assure if it is a winner or a loser? Thank you!

---

This depends on the payout level of the offer. You said you spent $30 bucks on UC browser. Not sure what GEO you picked, but if I remember it correctly, the payouts per install were like 30 or 40 cents - right? So 30 bucks could be considered enough to test a funnel with a decent amount of creatives, since its 100x offer payout. But its not enough to abandom the offer alltogether. Tho in this case, I would pass, app installs are pretty much done when it comes to whitehat promo methods.

---

It really depends on how much you're spending on testing each angle and traffic segment.

$30 may sound like a lot when testing a $0.30 offer, but how many different things were you testing? For example:

-How many angles?
-How many OSs (Android, IOS, Windows, Blackberry...)
-How many carriers and which ones? Are you testing both WiFi and carrier traffic?

For example, if you were testing 3 different angles across 3 different OSs for both carrier and wifi traffic, that would be 3x3x2=12 different combinations. Assuming you split that $30 equally across all combinations (which wouldn't usually be the case but I'm doing this to simplify the math), that would only be $2.50 spent on each combination - which may or may not have been enough to gauge promise, depending on how good the placements were.

If you can find one such combination that has significant traffic - for example 1 angle that works for android devices for a major mobile carrier in a high-volume-traffic geo - and scale that to multiple traffic sources, build up some conversion volume and ask for a pay bump...the revenue could really add up.

For all we know, this may be what your competition is doing. Then of course there are so many other optimizations they may have done that you can't see, such as optimizing bids, implementing day parting, and cutting underperforming placements.

There's really no need to be mad at yourself. Just because someone else has made the offer work, doesn't mean you have failed. It all depends on your approach. You may want to spend less money on testing each offer, but test more offers instead. It all boils down to personal style and experience.

UC Browser is such an old offer though. If you're a new person it would be harder for you to start at this point because you'd be competing with affiliates that have had many more months to spend on collecting data and optimizing their camps and are running on higher payouts - generally operating at higher profit margins and can afford to bid higher etc. However, there ARE new people that have made this offer work so if you want to give it another try, send me a PM and I'll put you in touch with the advertiser.


Amy

---

Thank you Matuloo and Amy.

Amy, I always feel amazed about how detailed your answers to all newbies' questions like mine. It will keep me busy all days if I was you. I am really appreciated.

There are 2 reasons that I will not bother you at this point:
1. Right now I can't guarantee that I can send high volume to the offer, so there is no point to get in touch with the advertiser.
2. Right now I'm with Clickdealer. My AM was so nice to me, I feel grateful to him so I will stick with the network as long as I can.

Thanks again Amy!

---

@andy2889 - One thing on point #2 - whilst having a good AM is great, you should always be split-testing between networks. Not split-testing offers between networks is one of the big causes of failure for newer affiliates' campaigns - the difference in conversion rates can be huge.

---

Thanks for that nice comment Andy! It takes good questions to get meaning answers, so thanks for your original question - I'm sure a lot of other people have the same question too.

I understand both your points and they're certainly valid. Feel free to take me up on my offer any time in the future.

As caurmen has pointed out, it's always good to split-test offers across networks. Your clickdealer AM will understand that profits what we're all after, and that if the same offer at another network converts better for you, that you'll naturally want to run with them.

Another advantage about running with multiple networks, is that you have more bargaining power when it comes to pay bumps. Basically you'd be in a position where the multiple networks will be "bidding" for your business. I've been able to achieve big paybumps just by being honest with 2 networks on how well the other was converting for me and how much the other network was willing to pay. I have a ton of respect for my AMs and a number of them have become personal friends, but business is business and should be treated as such, and everyone involved should understand this and not take anything personal (i.e. AMs should not feel offended about the fact that affiliates will want to do what will make them the most money; fortunately most of them do understand this).


Amy

---

Mr @caurmen, when you say to split test offers between networks you mean to run the same offer on the same geo, same carriers, sending them the same traffic? If that´s the case correct me if i´m wrong on this statement: I should not run them simultaneously, if not i´d be competing against me. I should, let´s say, run offerX-network1 today and offerX-netowork2 tomorrow. Does that sound reasonable?

Pablo.

---

@pablo You can rotate the offers at the same time( with your tracker, if you use one) without be competing against yourself and see which offer is performing better

---

Thanks!

---

What molotov said. The whole point of split-testing is to test several candidates for the same variable to see which one is better. If you don't run them simultaneously, you can't make sure the other variables are constant (e.g. conversion rate fluctuations over time of day, different amounts of traffic from different placements, different traffic quality due to changes in bidding position as competition varies, etc. etc.) As in any competition, the candidates need to be exposed to the same set of conditions. Otherwise the test wouldn't be fair.


Amy

---

Thank you guys for your advices. @Caurmen and @Amy, you guys are right. I feel like I don't have the power when it comes to negotiation. Whenever I ask for a bump, I only got a very small bump and my AM said"this is already nearly max-out..". So means 1-time-small bump only.

@pablo: Yes you set up rotation with 2 exactly same offers from 2 networks. I once test the exactly same Iphone 6s offer from Clickdealer (2$ pay-out) and Maxbounty (1.6$ pay-out). Guess what, I got higher EPV from Maxbounty offer.

---

@andy2889 - bumps are very dependent on what vertical you're running and what volume you're sending. In many verticals, once you've sent a few hundred leads you can get a quality check and potentially then a significant pay bump. Check out matuloo's recent post history - he wrote something about doing this in adult.

---

Yup, here it is : http://stmforum.com/forum/showthread...what-to-expect

Keep it mind its 100% about adult offers, but the general rules apply to any vertical more or less.

---

If you were an AM that's what you would be saying to affiliates too. After all, it would be in their best interest to maximize their margins as well - just like how we're wanting to maximize OUR profits.

I've heard that answer before - and then when I stopped sending traffic, the AM came back and offered me an additional 20%, which allowed me to scale the offer.


(((Applause))) Nothing like seeing actual stats to believe it really can happen. I was surprised too when I saw this for the first time - how could the same offer convert so much better on one network than another? But yup - it does happen!


Amy

---

I had a very similar experience when I started!

" Why should I test on several networks? They're all basically the same, right?"

"Oh, fine. People keep going on at me about it. I'll test, show they're wrong, and then get on with my campaigns."

"WTF? I just doubled my EPC by doing that stupid test?""

"Right, so, always test on a bunch of different networks, then..."

---

Standard rule: 100$ per test

You can make seperate rules and budgets for different types of verticals but if you stick to the standard rule, it takes away one less concern.

It will also give you the discipline and time to test out many verticals (if you are starting off) or offers (if you have picked 1 or 2 verticals you wanna work in) without spending too much money. Eventually you'll find winning offers, then its a matter of upscaling them as hard as you can as soon as you can.

---

Not sure such a rule makes much sense ... you want enough impressions, clicks and conversions so that you can get a statistically significant sample, not just spending a certain arbitrary amount.

$100 can get you 1000 clicks, 100 clicks, 10 clicks, 1 click, or no click depending on a wide range of independent variables.

---

Where do clicks cost 10$ or even 100$ per piece? I imagine that is gonna be very hard to get that profitable anyways.

There might be exceptions but the 100$ testing rule works great in 99.9% of all cases.

---

Haha, most of the clicks that we work with are well above $10

---

This simply cannot be true, from a number of reasons. I dont need $100 to test a $0.30 payout offer and figure out if it has potential or not. Especially not since I already have my set of benchmark creatives and know what to expect more or less.
On the other hand, $100 is absolutely not enough in many cases, Im pushing some PPS dating, with payouts around $85-$90 , how could I possibly make any conclusions in this case with a test budget of $100?

On top of that, when you are just starting you make so many mistakes that its absolutely pointless to set a rule like that.

This is to be handled on a STRICTLY case by case basis, there are no standard rules.

---

Definitely not.
They might work for a certain type of offers you have in your network, but $100 could be too much for a $0.10 APK for India and too little for a $28 pin submit campaign in Australia.

Payouts and type of traffic play a major role here to just generalise $100 is enough for 99% of the cases. Maybe it works 30% of the time. That could be a reasonable number.

---

Home > Paid Traffic Sources > Mobile