Post conference actions

Just a quick follow up on some of the advice from the shareware conference. The big pieces of advice I've made use of so far came from Dave Collins (Shareware Promotions).

Web Log Analysis

I downloaded WebLogStorming and fed my raw weblogs through it. This was really eye-opening, I found out that I could clearly track how many completed downloads I'm getting, tracing them back to a source or to search terms. I've been able to track basic things with the various web based stats services I've used, but the most important stat was never there.

The raw logs hold the bandwidth figure for each hit on the trial executable, so I can now tell how many times the download was completed.

Some interesting tidbits from the month of logs I had to analyze:-

• There were about 100 completed downloads from three download sites (not tucows or download.com). They probably aren't sales leads, as the names all imply "free", but I was surprised all the same.
• As far as I could tell, Yahoo sponsored search hits were hardly ever resulting in completed downloads. Comparatively, hits from MSN were looking like better value.
• Some of the more expensive paid keywords I've been chasing were hardly ever resulting in downloads.
• A surprising number of hits from Google's content network were resulting in downloads. This called into question the theory that this was wasted money, and suggested that I might have made a slight mistake in cutting back that spending (see below).
• Somehow, one of my images became the #4 google image search result for "american". That also explained some of the strange hits I've had from various unrelated forums.

One criticism of the software, it doesn't seem to do anything useful with xml/rss hits, so I couldn't really get a lot of useful info about my blog traffic from it.

If you've never used a raw log analyzer, I strongly recommend giving it a try, I don't think I'll ever want to rely just on web stat services again.

Taming Adwords

Dave painted a pretty bleak picture of Adwords in some of his presentations, he was convinced that if we weren't tracking our account carefully we would be throwing money away. He was right.

I hadn't really been paying attention to the split between Content and Search network, and I have the same campaign running on both. In October about half of our spend went on Content network clicks. I'd cheerfully increased the bid on a number of keywords, this had led to a bit of a surge in Content impressions and clicks. I spent a bit of time doing some housekeeping:-

• Cut the Content network completely from the main campaign, and added a second campaign for just the Content network. The Content bids are much lower. After realizing that I did get some useful traffic from the Content network, I've started increasing the bids slightly to get more coverage, but I'm being very careful and deliberate about it.
• I deleted some ad-groups with poorly performing keywords, and I made sure each group had at least two ads.
• Then I created a group for each set of keywords, and set the bids right back to the minimum. For the past couple of weeks, I've been making judgment calls when an ad stops showing, or when the position drops too low.

So far I seem to be getting more clicks for less money, and I'm able to keep an eye on the more expensive keywords. Also, the content network spending is much easier to track. I feel like I'm in control again.

Yahoo! Sponsored Search

I've added a tracking parameter to these ads to make sure I'm not missing any traffic, and I've trimmed my bids on some of the more expensive keywords. The fact that Yahoo! UK has a GBP20 per month minimum spend means I'll probably have to drop that soon. They can't give me enough traffic on affordable keywords to cover that, and the more expensive clicks don't seem to be worth the money.

Conclusion

I should have done all this months ago...

Tags: eswc06 micro isv adwords