The question is the topic. How do you get valuable date from your failures? Specifically targets, angles, placements, offers etc. It's pretty easy to see when you mess up your tracking links or testing your campaigns. Those things are black and white. Make an adjustment and move on.
How do you comb through all the data of failed campaigns and come out with some concrete causal-factors? I've been trying to do this, but it's been more of a gut feeling so far. I haven't gotten a decent system in place.
Any thoughts?
Oddly enough this is something I went through a while ago.. not so much from targets or placements.. more so from offers and angles, and matching them with traffic sources.
I took a list of my last bunch of failed mobile camps/offers.. where I'd ran them, angles I'd tested etc..
Looked for a common theme between them all..
The main common theme I found with some of mine was.. I wasn't researching the traffic source properly before matching it to a certain offer. For instance.. Buzzcity, has far more volume for feature phone users in most developing countries, than anything else..
I took extra time to find some offers, for the countries I knew had conversions in them, with offer landing pages that would sit correctly on a feature phone. Then built the bridge between that and my banners.. worked out ok, but had to look elsewhere for better traffic after a while.. but I found it useful doing this..
That's one thing that I haven't done very well yet. Is really look at the traffic sources and what they run. Makes a whole lot of sense to match the offer with the traffic, but I've just been using a shotgun approach rather than a rifle. I'll definitely keep that in mind!
The three things I'd tend to look for in a bunch of failed campaigns are:
1) Things I'm doing the same every time. It's very easy to think "I've tried everything" when actually you mean "I've tried everything... except that I always used the same font, my framing is pretty much identical on each image and I'm sticking to bold primary colours with a 2px wide border". Once you know what you're not doing, you can try that too.
2) Overall trends - as dynamicsoul says. VERY useful.
3) Outlying parts of the campaign / campaigns that did work. One image that got a high CTR, a landing page that showed a noticably higher conversion rate, etc. Match these up with other successful parts in other campaigns, and suddenly you've got a winner.
It's quite easy to spot similarities in winning creatives, or in the case of a failed campaign, the 'best losing' creatives.
I store an archived library of banners and landing pages for every country I advertise in. This isn't as simple as copy and pasting every creative you ever use in to a library. You have to actually maintain it, weeding out the poorer performing stuff as and when you find it.
No point in storing a massive collection of fail.
But it took me a surprisingly long time to realise that.
I use folders in Evernote (https://evernote.com/) with a file for every single traffic source that I work with.
These are jam packed with existing advertisements from competitors, links to their landing pages, notes on the offers they run. I'll cross-check what I can with tools like WRW to nail down longevity/freshness.
Before I spend any money, I'll reach out to a rep at the traffic source and get an idea of the traffic/demographics/etc, then store all of this in Evernote (as well as printed copies). I'll print out my transcripts of Skype convos/emails etc to store in a paper file. Good for reading away from the desk where I'm less distracted.
I'll also comb forums and blogs for snippets of info that could be useful. Only then will I start actually sending traffic.
Extracting data from the failures is a combination of intuition and, well, bothering to check the data. Something many affiliates neglect in favour of basic profit/loss.
When I do my morning stats, besides recording profit/loss, I take notes on the performance of the landing pages and offers. Usually on Mondays or Tuesdays I'll compare week-by-week performance to see if I can spot trends, or signs of decay in winning ads. I'll also compare what I've actually been doing to what my initial research told me I should be doing. (Critical)
If I'm close to getting profitable on a platform, I'll record what I consider to be the 'benchmark' figures of success. I.e, what CVR, CTR and offer payout do I need?
Even if I can't get there today, it's important to store the attributes that would turn the campaign in to a success. If three months later, you make a breakthrough on a different traffic source with a different offer, you can look at your benchmark figures and quickly determine if an abandoned traffic source is worth another shot based on the improvements you've made across the board since you last attacked it.
Experience does play a major role in learning from your failures. But keeping archived records of past performance is just as important, as is taking the time to actually look at them instead of diving balls deep in to whatever catches your eye.
Finch would you be willing to take a screenshot of your organization of all of this. I'm just starting to run campaigns and would like to start doing this from the beginning
wow, all I did was keeping a journal of what I did, what worked and what didn't + having folders for all the ads I run...
Finch blasted my ass to a whole new level of organization!!!