Is it the affiliate's responsibility to improve the loading time of banners and hence keep the banner file size smaller (and probably compromising on quality, although not sure if the reduction in quality in a 320x50 banner would be perceivable but still assuming it's perceivable) or do we not need to worry about it?
BTW is any particular file format for banners preferred? Right now I'm using gifs.
Man, here is what you need to know:
In this business, EVERYTHING, and I mean EVERYTHING is affilaite's responsibility.
You would think that your network will inform you when offer page changes or if the offer page goes down: THEY DONT
You would think that your advertiser would help you with weeding out the bots and getting what you are paying for: THEY DONT
There are people who do this, but you cannot expect them to take care of your money.
Now that I have found a place to rant, let me get to your question:
- Use this: https://kraken.io/web-interface.
- Try to keep file size under 15kb for 320x50 banners.
- Gif is fine.
- here: http://stmforum.com/forum/showthread...late-inside%29
Good luck mister!
So here is what I did, I had all of my banners in .gif so I converted the ones which didn't have any flashing/blinking text (animation) in them to jpg using the default conversion settings on http://www.zamzar.com/ and then ran the converted .jpg through Kraken with lossless (didn't choose lossy because the degradation in the sharpness of the text in the banner was perceptibly compared to the .gif). This resulted in me ending up with ~10KB banners in .jpg instead of ~15KB in .gif with not much loss in quality. Is this the right way to go about it? What do you guys use for converting formats, Photoshop with various quality settings which you obtained by tweaking?
However I still have a couple of banners with blinking text in .gif which are 18KB and 30KB. I see no way to reduce their file size (even Kraken didn't compress them). Ideas?
that's interesting... With which tool did you create your banners? I made mine (quite standard as per the post from Stackman that arun suggested) and they end up with something like 4 to 7 kB which then is reduced by approx. 20% on average (not much, but hey, every kB counts!)
Have you guys read this case study ?
This thread reveal again that everything have to be tested and the results can be surprising and you can actually find something to use as your advantage (hint)