Mobile Cookbook: If You're Having Trouble Getting Traffic Fix It Like This (9)
Mobile Cookbook: If You're Having Trouble Getting Traffic Fix It Like This
I've seen quite a few people having problems on Decisive getting traffic for some of their campaigns, so I thought I'd write up a quick guide to smoothing out traffic problems.
Decisive's traffic algorithm is pretty sane as these things go, but like most traffic sources it has its quirks.
So for everyone stuck with a $0.20 spend per day on their carrier campaigns, here's how to uncork the traffic bottle!
Pushing Up Your CPA Goal
Firstly, it's worth looking at your SmartCPM CPA Goal.
If you're not getting any traffic on a new campaign, your CPA goal is likely a bit low for the cost of the traffic. The CPA goal guides the system's dynamic bidding, and if there's simply no traffic that it thinks will be worthwhile, it won't bid. Bump it up to 2x or even 3x your payout (in extreme cases, even 4x payout) and see if you start to get some traffic flowing in.
If your campaign's not getting a lot of conversions, and traffic has dried up, this is thanks to the Smart CPM optimisation system. The Smart CPM system tries to minimise your spend, and if it "thinks" the campaign isn't working, it will start shutting down - reducing bids lower and lower.
To counter that, you've got two choices. You can either raise your CPA goal again to give it a bit more room to breathe, or you can switch to flat CPM (see below). If you want to continue firing traffic at a campaign that's performing at significantly negative ROI, a flat-CPM bid is probably simplest.
Raising your Daily Limit
If you're pushing your bids up and up but the traffic's still dribbling in, then you should also take a look at your Daily Limit.
The Decisive Daily Limit algorithm is extremely conservative. Wheras other traffic sources take your daily limit as somewhat approximate and will often spend over it, Decisive's system is set to almost never go over your limit. Unfortunately, that means that a lot of the time it will dramatically underspend rather than overspend, particularly in geos where you're not in great danger of spending $1000s per hour.
In my testing, I've found that in low-volume targets (say, Site + Carrier + Android in Brazil, a reasonably low volume option with a lot of competition), the Decisive algorithm will aim to spend about 1/3 to 1/4 of your Daily Limit. So if you've got a Daily Limit of $10 set for the campaign, it'll aim to spend about $3 - or about $0.12 an hour. It'll go higher at peak times and lower at quiet times, but that'll be about the average.
If you want to get data faster than that, you'll have to raise your Daily Limit. This is obviously a bit scary because you don't want to overspend, but it's the only way of avoiding throttling on these low-volume campaigns. The 1/3 to 1/4 rule appears to apply until you're eating all the traffic you can in smaller geos, so if you start by setting your Daily Limit around 3x what you'd ideally like to spend, you should see a good balance of spend limiting and traffic.
DO NOT do this in high-volume geos. If you're targeting US / Wifi / all handsets, you really want your Daily Limit set to what you want to spend, no higher! Generally, you shouldn't adjust your Daily Limit until you see that you're undershooting.
Flat CPM
If Smart CPM's giving you headaches, I recommend switching to regular CPM. It doesn't do the auto-optimisation thing, but you should be optimising placements manually anyway, and it's a lot easier to work with.
Decisive's system responds in real time to bidding changes - I've been testing this today - so you can very rapidly test the volume you'll get from flat CPM bids. Just record your daily wins for that campaign, adjust your bid, set a stopwatch for 5 minutes, and then at the end of those 5 minutes record the new daily wins - that'll give you a decent approximation of how much your bid change will affect your available volume. Assuming you have a reasonably sane Daily Limit, you can bid quite high even in very high volume geos and get a good idea of the volume without overspending massively. I'll usually bid around $2 CPM, sometimes higher, for these 5 minute tests at least once just to get an idea of the total volume available.
If you're nervous, keep an eye on your realtime stats and hammer the Pause button if it starts getting out of hand.
Remember "Legit" Verticals Get More Traffic
A lot of Decisive's advertisers aren't too keen on some of the more popular affiliate verticals - notably Antivirus, Sweepstakes, Diet and most Dating.
Campaigns that are outside that list of verticals have access to additional (possibly higher-quality) traffic sources.
In my testing on a limited subset of geos, I found that the amount of traffic available for campaigns outside those verticals was about double what campaigns in those verticals could access. But I strongly suspect that varies a lot by geo and type of traffic - test to be sure!
Obviously, run the vertical that works for you! But if your traffic is low, be aware of this factor.
And there you have it! If you're having volume problems, check your CPA Goal, raise your Daily Limit, switch to flat CPM and test out different bid types. And that should sort you out!
I hope that was helpful! Any other comments, suggestions, or tips on solving your traffic woes? Post 'em below!
Would it be a good idea to look at the overall CPM of the campaign of a Smart CPA campaign and then try to run a flat CPM campaign at that same CPS rate? For instance, I might have a Smart CPA campaign with a goal of $1.45, but looking at the dashboard, I see that the average CPM for the campaign is .23 and I am winning 20% of the bids.. Could I theoretically clone the campaign and run it flat CPM for .23 and win that same 20% of bids?
"Remember "Legit" Verticals Get More Traffic" - does it apply to other sources as well?
@mr_mac - I don't know for sure but I would suspect it applies to some other DSPs too, yes.
@fishinseo - probably not, I'm afraid. If the campaign's been running for a while, the chances are that the Smart CPM is bidding on different placements very differently. You'd probably need to drill down a bit more and run campaigns targeting small placement groups to get exactly the same performance - that's the disadvantage of moving away from SmartCPM.
Thanks for this post! Tried with flat CPM the other day on decisive and I had to hammer that stop button because it started coming in like crazy!
Trying #2. Raising your Daily Limit
I add +300% of daily campaign budget for testing new geo
Here is the information I got from Decisive supporter when I asked him the same question about how to get access to higher quality placements with the system:
'Neat' means if you are promoting for e.g. ABC Game, the creative and landing page should explicitly say it's ABC Game (no catch or tricks). Of course, you can be creative in your own way.
Antivirus, other utilities, sweepstake offers and dating campaigns are not allowed on most of the good traffic.
What about go2mobi? i.e. when bidding on 320x50, one big carrier and one geo I don't get *that* much traffic. Would think there would be more, running 4-5 exchanges and setting bids per exchange when starting.
@jessejames - how does the traffic you're getting compare to the traffic they show in their inventory tool?
Home > Mobile >