Welcome to Curious Business
Every Friday, I post a small insight into running Curio City and/or Blue Hills Editorial Services. My most recent posts are directly below. You can also start with the first post, or use the subject labels to the right to home in on particular topics. Feel free to comment on anything that interests you.
Friday, August 19, 2016
You Know It's All About Those Bucks
I finally decided that I really did break Google's conversion tracking when I tried to improve it.
If you remember this background from two weeks ago, you can skip this paragraph. My old tracking code tells me when a particular keyword "converts" to a sale, and I can see how much I spend on that keyword per conversion, but I can't see what the conversions are worth. The new script would report that information so that I can allocate my ad budget a little more intelligently. Paying $15 to convert a $20 sale is bad; that same $15 expense for a $100 sale is fine.
If I'm reading my test results right, the "new" script is capturing the value correctly but not reporting conversions to Google. I reverted to my original files once I was pretty sure that I'd broken that, and the conversions started racking up again.
I sent Google a detailed question with supporting files and got my first unhelpful answer a week later: " I have checked your account and found that the conversion code that you have added is working fine and tracking conversions." He attached a screenshot showing the Tracking Status as "receiving data." Fine, but that was after I had reverted to the old files (as I had, of course, already told him). I asked him to check the week of 8/1 - 8/7, when I was using the new files. A few days later he replied " I have checked the conversions from Aug 1st to Aug 7th and see that the conversions have been recording." and attached the same screenie. Fine, but his results showed only ONE conversion using the default value of $20 (from the old code, not the new code, in other words).
Fine, I can plod as slowly as the next guy. On Tuesday I uploaded my broken "new" files again and told this fellow to check back in a week. By then he'll see no conversions and the Tracking Status will show no incoming data. Once I prove the problem exists, we can finally begin troubleshooting the code that he should have been looking at two weeks ago.
His last reply came yesterday, asking to turn it into a telephone consult. Uh, no, not going to do that. This is a technical problem that needs to be solved in writing, and the telephone will only introduce more confusion. I know this is just some kid in India following a script, but maybe it will get passed along to a real technician whose first language is English after I can prove that there's a problem.
Unless that happens the odds of getting an accurate answer or fixed code are very low. The sensible thing to do would be to simply drop the matter and stick with my old files. But now that my appetite for better data is whetted, I'm going to keep at it, even though that will mean missing a week's worth of conversion data just to prove that it doesn't work. Of course, I need a meaningful number of sales during that week in order to generate data. I can't prove anything if it stays as dead as it's been for the past few days; in fact, there's a small chance that the file is working properly and simply doesn't have any valid conversions to report.
I should state for any nervous readers that this is all anonymous as far as I'm concerned. Yes, retailers have the ability to track you all over the internet, and most of the big boys do. I don't. I only want the keyword statistics.
In related news, my developer is getting ready to update Sunshop and make a couple of other minor technical changes for me, including re-implementing the newsletter signup dialog. If I can get all of this smoothed out by the beginning of September I should be able to squeeze an extra percentage or two out of this Christmas.
Labels: Reasons to hate Google