Home > Paid Traffic Sources > Mobile

Question about Optimising the Campaign Vs Day or budget? (9)

07-20-2015 03:45 PM #1 mobybyd (Member)
Question about Optimising the Campaign Vs Day or budget?

Hey Guys,

So I have been running a campaign for the last 4 days.

Got the data on Day 1 Then I have Optimised it for Day 2... Day3 Saw increase of then I optimized further using Day 1 + 2 Data and so on to increase the ROI.

However I m kind confuse on which day data to use because i dont see stability on ROI ..day 3 was 75% roi and other day 4 it was 1% Roi.....My question is:

Which data should i use to optimise on the current day ...The data for the yesterday only? the last 3 days or last 7 days?
Or should it be per budget/day ..which is the same...

Thanks
Raki

07-20-2015 10:57 PM #2 jjbachiller (AMC Alumnus)

I'm pretty new too and english is not my main lang so don't have my answer too seriously but... you have to find the pattern of your data. Perhaps the ROI change between days but it's very similar in the same day of each week (for example between 75-85% on Mondays and 1-3% on Tuesdays). You need more data to find the pattern and optimize properly, don't force the optimization or you ended with a very little amount of traffic.

07-22-2015 06:08 PM #3 mobybyd (Member)

Hey

Thanks for the answer!

So you think I should wait 1 week before starting filttering and see a patern?
One of my campain had a positive ROI for After Day 2 when I start optimizing, however from Day 4 5 and 6....it was on negative...
Should I reset it ? Kind confuse..

07-22-2015 08:10 PM #4 vortex (Senior Moderator)

May I ask what it is that you're trying to optimize for? Offers? Banners? Landers? Placements?

Some traffic sources provide the option to spread traffic throughout the day. It may be good to do that to minimize the effects of fluctuations on your data.

Of course the longer period of time you collect your data for, the more accurate you'll be when making decisions. However, that may require more test budget as well. If you're just comparing 2 similar offers for example, it may be better to just ignore the fluctuations and cut as soon as the data reaches statistical significance.

It's always possible to shoot for more accuracy, but the question is "how much more benefits will you get for the additional cost?" And my guess is that everybody's answer will be different.

Amy

07-23-2015 12:41 PM #5 mobybyd (Member)

Hey Amy! Thanks for the link...it helped i lot.

I started by optimising the Banners but i did i mess up, so I checked the link you have provided and applied the optimisation to the Banners - it working now

Then Im planning to optimise the Placements ( which app/site bring best Cpa ) - I dont have a lander, I direct link

What do you mean by or how do you optimize the offer?

Thanks!

07-23-2015 04:25 PM #6 arunbasillal (Member)

If a campaign is almost profitable or breaking even right off the bat, I think the best strategy is to optimize only the stuff that you know is not going to change in the future.

Say a banner has 0.02% CTR, after day 1, that needs to go. Or if a placement has 5000 clicks and zero conversions. The obvious ones.

Other than that, I would run it for a week before optimizing too much. Especially if its breaking even or profitable right off the bat.

07-23-2015 09:06 PM #7 vortex (Senior Moderator)

What do you mean by or how do you optimize the offer?

By that I meant test similar offers - in the same vertical and geo - and even the same offer available on different affiliate networks.

Then you'd just compare all the offers and cut using the split-test calculator until you're left with the best one.

Amy

07-24-2015 04:00 PM #8 mobybyd (Member)

Hey !

When you do you guys optimise for the variables ( carrier / Phone Brands / ISP etc...) ? After you cut the banners / placement etc..? or Before or in the same time?

Thanks
Raki

07-25-2015 04:47 AM #9 vortex (Senior Moderator)

Originally Posted by mobybyd

Hey !

When you do you guys optimise for the variables ( carrier / Phone Brands / ISP etc...) ? After you cut the banners / placement etc..? or Before or in the same time?

Thanks
Raki

It all depends on how optimized your other variables are. For example, if you're using banners or landers that don't convert well, you can't really blame placements or carriers for not converting. Having said that, it may be a good strategy to temporarily cut carriers, placements etc. that aren't converting well currently, and just use the best carriers and placements etc. to test and optimize creatives first. That way you keep your testing as cheap as possible. Then, when you have some good creatives, you can go back and retest carriers, placements etc. to give them a fair assessment.

Of course everyone will do things differently but this is how I do it.

How do the rest of you do it?

Amy

Home > Paid Traffic Sources > Mobile