When talking to customers about web slices, we touch on several important topics. Naturally, we always talk about adoption, discoverability, content relevance, making the button light up, and analytics. Yes, you want to leverage our gallery, which has millions of users visit each month, so that others will find you. But, you also want to make the slice easily found on your site. Your customers are the easiest first adopters. So, let’s build you an audience.
But, how would you know that a web slice is successful? This could be a complex question, especially when first considering web slices. After a customer launches a web slice, they’ll ask “how many installs did we have on the IE gallery?”. And while important, one has to consider the install numbers from our site as well as your site. But, there are challenges in tracking install events. We can get the numbers of people who visit your gallery page and even how many clicked on the “Add To Internet Explorer” button. However, that’s still not accurate, because folks can install from your side too. Then, there’s another issue.
One weakness in our web slice story is that IE doesn’t expose when someone installs via the Feeds Button (command bar) or in-document discovery (frame in the page). The only way to track installs on your site is via a control that your developer places on the page (like a clickable image or button, where they call AddToFavoritesBar to install the web slice). And because of this, the best way to understand your current audience size is to leverage your analytics engine (and you better have one today).
To get the most realistic view of web slice usage, there are two metrics to monitor (talking in terms of the 3 page pattern):
1. Total Subscribers – track the traffic on the update page for how many clients are polling your update source (FeedURL).
- This will tell you how many machines have your web slice installed (i.e. they have your web slice button on their browser).
- If you do a 1 or 2 page pattern, it’s a different method – you’ll have to filter on the agent string for the requests generated by the Windows Feeds system.
2. Total People Viewers – track the traffic on your display page (assuming Alternate Display).
- This will tell how many clients are actually looking at the web slice, and this should be a subset of the first number.
- You can do normal analytics on this page, and apply your normal thought process of making conclusions on the UX.
- If you’re doing a 1 page or 2 page pattern, tracking can be done but you have to differentiate a visit to the page, the polling for slice update, and the web slice display request.
In summary – the first point is your total current installation base, and the second represents how many people are clicking on the web slice button. Both numbers are significant independently, but when viewed together you get a sense of reality.
Potential conclusions (subjective):
If Total Subscribers (#1) is flat or decreasing, then you have some type of problem.
- If flat, are people not aware of the web slice? This is why we encourage your marketing of the slice, and sometimes this involves creating awareness / education.
- If decreasing – are people removing the slice for some reason? Consider reviewing UX, content relevance, update frequency, etc. ?
If the ratio of Total People Viewers (#2) to Total Subscribers (#1) is going down (or low), then that is more tricky. My bet here is that deleting is not far off, which will start impacting Total Subscribers (#1).
- The question is: why would a user keep your web slice, but not click on the button and not delete it?
- Would a changing title in the web slice button help?
- Is it a problem with the consumer not appreciating the content or look from a previous visit?
- Take a review of the IE8 Web Slice Style Guide to see if you agree with some of the suggestions, especially Appendix sections.
Your web slice is another web page (or two), so use your existing skillsets and analytics engine to find out what is happening. And don’t just ask about installs, ask what is the size of your existing web slice audience and start tracking the trend. Hoping this helps in understanding how your slice is performing.