CHEATSHEET: Do's and Don'ts for Mobile Display (12)
Looking to get started with mobile display (banner) traffic? Here are some tips that can shorten that learning curve.
********************
DO's:
DO consider testing multiple offers from the start to increase your chances of finding a good one. Also test same offer on different networks, and different offer landers for the same offer.
DO test low-competition geos first. Consult an ad planner for geos with lower bids. Asia, Africa, & LATAM America are gold!
DO test lots of angles. Can't make a campaign work without a good angle!
DO test multiple banners per angle. Testing different headlines and images for each angle will give you a much better picture than just 1. e.g. 3 banners/angle to start.
DO optimize your lander for speed. Especially important for mobile! See tips here, here and here.
DO get rid of bad placements based on conversion rate. See here for a helpful tool.
DO consider cutting bad placements based on CTR. See the end of this post for suggested guidelines (thanks sebastian_r for the reminder!)
DO test direct-linking vs. landers.
DO consider using "generic" banners to quickly test offers that have broad appeal. Stuff like "Download Now" and images of blinking presents / envelopes, "you've got a message", etc. Some traffic sources and exchanges still allow these. (Thanks sebastian_r!)
DO make sure there's not too big of a "gap" from banner -> lander -> offer. Visitors need to know how the offer will solve the problem stated in the banners and/or landers (thanks wiifmdude).
DO target different banner sizes. 320x50 isn't the only size available! Just make sure the sizes you're targeting have enough traffic.
DO test wifi and carrier traffic separately. Because they can perform very differently.
DO test app traffic and mobile site traffic separately. Same reason as above.
DO identify most-used handsets from your campaign stats make sure your landers display correctly in them. Use for example Google Chrome Developer Tools or BrowserStack.
DO test PNG banners, not just JPG. See here for details.
DO consider customizing your creatives to certain types of placements to increase conversions. See how hlyghst did that here. Make sure there's enough traffic!
DO consider testing on pop traffic first. Once you have a good offer and lander, can test display and perfect your banners (see sebastian_r's post for explanation).
DO verify lead quality before scaling. If all is well ask for a pay bump.
DO use traffic network tracking tokens to help you optimize your campaign. You can find tracking tokens here (thanks to mitesh muley).
********************
DON'Ts:
DON'T test everything with a single campaign. Wifi/Carrier/App/Site, different banner sizes / ad formats, different geos.
DON'T assume all or even most of the clicks you're getting are real. Bots are rife on mDSPs.
DON'T be afraid to bid high. That's how you get the best quality traffic.
DON'T set a random bid. Consult mobile ad planners or ask the traffic source rep for bid recommendations.
DON'T forget to set a frequency cap, at least in the beginning. 1 in 24 hours is a good start.
DON'T be surprised if interstitials give huge CTRs but low ROI. Reason - lots of accidental clicks. However, sebastian_r tells us that carrier traffic can work well for high payout pin submit offers.
DON'T be cheap when it comes to getting proper translations! Use OHT for speed or Fiverr for cheap pricing.
********************
Mobile Display Sources to Try:
Need more? Check out these threads:
Android Ad Network SDK Stats
EPIC list of 400+ Mobile Ad Networks
********************
Great Posts on Mobile Display:
Mega-infodump of STM's mobile guides, tips and case studies
sebastian_r: Various great tips on mobile display
Amy
Reserved just in case!
Good one!
Here are my biggest take aways from display:
Cutting twice is necessary when running on CPM. Need to cut placements which get impressions but don't deliver clicks. Need to cut placements which deliver clicks but no conversions.
App traffic always on auto block. Could never make it work with worthwhile scale.
Exception, interstitial carrier can work very well with high payout pins, but the click loss is brutal.
I'm always going from pop to display. If something works on pop, I scale to display. Removes one moving component in the initial testing and pop is able to deliver lander impressions for split testing quicker. At the same time, there are many verticals which can work on display but never on pops. So going from pops to display is not always feasible.
Important: Display users are coming from a different environment than pop users, its not necessary to be as aggressive as on pops. Can even be counterproductive. Keeping camps alive on exchanges like smatoo and mobpub is a tough one when running aggressive.
Too lazy to create banners and running an offer with wide appeal? Good old Download / Play still gets around >1% crt on many placements IF you get it approved.
Originally Posted by sebastian_r
I love using generic banners to quickly test offers as well. Basically that's the closest to turning display traffic into pop traffic we can get. Too bad not a lot of mDSPs/exchanges allow them anymore.
Thanks again for contributing your valuable insight - very much appreciated!
Amy
From the low responses to this post, seems that display is a bit out of fashion.
Big opportunity right here.
Guys,
Any particular test time from or a test budget after which you definitely know the offer will not convert on pop/display specially for App/Mobile content campaigns ?
Best guide, ever on mobile display.
Display can be a tough one. It makes sense to do the research first and find suitable apps that will suit your demographic and has volume available. Make sure your lander and offer "fits in" with the particular theme/style of that app for best results.
While I agree with Sebastian that app traffic is generally poor, I have made it work with select offers when thought about and planned properly.
Hlyghst (I think thats his username anyway!) once revealed he targeted an iphone 6s sweeps offer at users of a weather app - His banner warned of an incoming tornado and used an image of a "Phonenado" and pocketed some nice profit by thinking outside the box
Very nice guide! Thanks Amy.
On Google Display Network, be sure to exclude adsenseformobileapps.com and also set a mobile bid adjustment of -100%. It is not recommended to mix mobile/desktop as they tend to behave 100% differently.
Originally Posted by nimish
If it's pop traffic, and if the vertical is a really popular one where ripping landers is easy, I would just rip a bunch of the most-popular landers, launch 3 identical camps at low-mid-high bids, run maybe 10-20x payout traffic to each. If the offer doesn't convert at all then there would be little point to continue. (However, I usually test multiple offers for a geo+vertical to give me a better measure of how easy it would be to make that geo+vertical profitable. In that case I would run more traffic to the camp - out of the 3 - that gives me the best ROI, and then see whether any of the offer+lander combos are close to breaking even. I would also drill down to OS -> Carrier to see if there are traffic segments that converted well, that I might test further on.)
For display it's more complicated, because now we're needing to test banners also. The angle is everything when it comes to banners. I would do my best to come up with a lot of angles, pick what I think are the best 10+, make 3 or more banners per angle, and run enough traffic to see whether the angles show promise. If none of the angles are getting a decent amount of conversions (I'd say -50% ROI - although that would depend on whether there are bad placements that have drained the budget without converting), I would usually give up at this point and test something else.
Exception: Say it's a relatively new international offer that a lot of people are having success promoting, then in that case it would be worth it to keep testing angles until finding one that works, and go from there.
Sorry if I'm all over the place - but hopefully I've addressed some of the factors that need to be considered.
Amy
Nice write up! Thanks Amy.
Home > Paid Traffic Sources > Mobile