Is Google Conversion Tracking Accurate

[Mathese]

I recently implemented Google's conversion tracking on website. After a few days of data, the results do not match what I have for sales from my internal tracking. I am concerned that one of the methods is not working and I am assuming it is Google. I do not know enough about coding to understand exactly how it works but I know it basically counts a conversion each time some hits my thank you page. However, this does not seem to really be the case.

Is anyone else finding Google to be inaccurate for their conversion tracking?

[lizcamps]

Hi Mathese,
That's an important question which I'm surprised more people don't ask.

In my experience the accuracy of the Google conversion counter varies widely when my clients' IT departments handle implementation themselves; whereas it appears quite accurate when we are in control of site production and log analysis.

I recommend to any readers that if someone else is placing the code for you, that you conduct a few days' worth of extensive tests to ensure they did it properly. Not to diss any client, but when they implement the Conversion Counter, they've never done it before. Having experience in placing this code does make a difference in accuracy.

But in your case, Mathese, you seem to say you did the implementation yourself, rather than rely on a third party. So there has to be another cause for the conflicting metrics.

NOTE: If you have a dedicated Google rep, turn to them first. If your spend is high enough they will offer to have one of their techs advise you on your implementation. The more info you can give them (to save them time), the faster they can get back to you with actionable advice.

If you don't have a rep you can try to get one (easier if you spend over $5K/month), and/or rely on your own devices. Some troubleshooting suggestions:

- Make sure the code was not placed inside an include.

- Analyze the logs and isolate the stats for you thank-you page. Are the incoming pages that lead to your thank-you page any other pages than you would expect? In other words are there any methods of getting to your thank-you page, other than the shopping cart? If so, you'll have to block that illicit flow.

- Are there any "No Referrers" coming to the thank-you page? If so, there are probably people bookmarking your site from the thank-you page or using their History file which may show that page as the most recent, so every time they revisit your site they are throwing off the conversion data.

- Don't combine multiple types of conversion trackers (e.g. thank-you page, plus a "page view" counter on every page of the site). The need to disentangle multiple snippets of Google code on each page makes it harder to
solve the problem.

When I've brought up the issue in the past with my Google rep, she had recommended that we always implement the code ourselves (I agree since we're so familiar with it), and that we use Google Analytics as it includes Conversion plus much more data for the whole site. I would add that you should use a log analysis application to cross-check any Javascript metrics, just as a routine best practice for any site.

Mathese, it would be helpful to know whether your Google conversion metrics are over-counting or under-counting sales.