Awesome that they are open sourcing this. I've always felt funny about paying a ton of money to FullStory for this. I wonder what this means for FullStory and HotJar?
I expect they already offer more and better differentiation to not be too terrified of this. I work for an analytics company and when our core features are copied by larger companies it’s pretty exciting and is a driver of innovation for ourselves.
It's nice to know how your users use your apps and web-sites, but I think it's morally wrong to track user interactions in such detail. User testing is more effective and they know they are being watched. It's probably a good idea to show a disclaimer that says everything the user does on the site will be recorded, or better, have the user opt-in to being recorded.
I work for an analytics company - that said, their malware example is one unlikely to come up in most user testing. I say most because I’ve seen some do user testing over Skype with the user’s real environment.
Another piece of anecdata, I had a customer experiencing a bug I am confident I wouldn’t have uncovered even with intense technical communication without having session replay... our UI was breaking a POST request for a user and there was seemingly no way for it to do so, it turns out in a number field, Firefox had these up and down arrows to move the number up, but the attribute that received the value when doing this was different from expected when entering the number by keyboard.
Anyways, the scary side of session replay does arguably outweigh the upsides. If you don’t know or care, and even if you do care it’s hard to make sure you never capture things like credit card numbers, etc...
At my current company I’ve helped push a lot of things that make it darn near impossible for us to capture PII, but nothing is perfect.
As someone else in this thread said, it would be nice if someone built an end to end OSS analytics tool that was anonymizes by default, but I think even the concept of anonymization is hard to tackle depending on how one defines anonymous. The set of features one uses, and the frequency and/or order of use can likely be identifying by some definition.
Opt in, for the example I described (bug hunting) would be a totally reasonable use, but I bet you’d have a hard time justifying the cost of such a tool if you use it for such sparse cases, and you lose out on many other serendipitous discoveries. Again, if you think it’s morally wrong, there’s not much to argue with in terms of using such a feature :)
OSS analytics packages are unfortunately such a rarity - Piwik seems to be the only serious game in town. All I really want is an anonymized picture of feature uptake and some inkling as to how well we're doing in the UX and perf departments. I want to have these things without that annoying EU cookie widget, without shipping any data to third party services, and without loading any third party scripts.
Unfortunately this seems to just be the frontend tracker, and the backend will remain a closed MS service.
Snowplow[1] is an alternative to this. While they have a managed solution, their backend is open source and can be self-hosted. And they have a forked version of the Piwik client side library designed ot work with Snowplow[2].
Snowplow is a bit more generalized than just Piwik, and it shows in Piwik having a more robust featureset for website analytics specific. But Snowplow has a lot more usefulness if you're having to merge a bunch of data sources together to get a picture of what's happening.
How does clarity differ from other tracking or session replay tools like hotjar or fullstory? It looks like it's another Microsoft-ish-ly bad version of other existing things.
It's not louder, at least doesn't sound it to me, but yeah the background music should be lowered just a bit more. I'm surprised the guy isn't yelling as he talks, felt like I was watching something out of Rick & Morty, one of those awkward ads.
Creepy. Notice how it doesn't just record where a visitor had scrolled to or has clicked, but it apparently sends either DOM content or actual screenshots to the server. Otherwise they would not have found out that the malware put extra stuff in the user's Bing results.
When Manali started using Clarity, she realized that a lot of users were abandoning the page before reaching the bottom, which has important information about the recipe. After replaying sessions of users who abandoned her blog, she noticed that users who were only interested in the recipe, which is at the bottom, scrolled through the long post and gave up midway and abandoned the page.
They should find a more compelling example. "Recipe at the bottom of long page with life story and huge images" is one of the canonical examples of user-hostile web design. EDIT: As in, no one should need a product to tell them this is clearly not what people want.
16 comments
[ 2.9 ms ] story [ 43.9 ms ] threadAnother piece of anecdata, I had a customer experiencing a bug I am confident I wouldn’t have uncovered even with intense technical communication without having session replay... our UI was breaking a POST request for a user and there was seemingly no way for it to do so, it turns out in a number field, Firefox had these up and down arrows to move the number up, but the attribute that received the value when doing this was different from expected when entering the number by keyboard.
Anyways, the scary side of session replay does arguably outweigh the upsides. If you don’t know or care, and even if you do care it’s hard to make sure you never capture things like credit card numbers, etc...
At my current company I’ve helped push a lot of things that make it darn near impossible for us to capture PII, but nothing is perfect.
As someone else in this thread said, it would be nice if someone built an end to end OSS analytics tool that was anonymizes by default, but I think even the concept of anonymization is hard to tackle depending on how one defines anonymous. The set of features one uses, and the frequency and/or order of use can likely be identifying by some definition.
Opt in, for the example I described (bug hunting) would be a totally reasonable use, but I bet you’d have a hard time justifying the cost of such a tool if you use it for such sparse cases, and you lose out on many other serendipitous discoveries. Again, if you think it’s morally wrong, there’s not much to argue with in terms of using such a feature :)
Unfortunately this seems to just be the frontend tracker, and the backend will remain a closed MS service.
Snowplow is a bit more generalized than just Piwik, and it shows in Piwik having a more robust featureset for website analytics specific. But Snowplow has a lot more usefulness if you're having to merge a bunch of data sources together to get a picture of what's happening.
[1] https://snowplowanalytics.com/
[2] https://github.com/snowplow/snowplow-javascript-tracker
They should find a more compelling example. "Recipe at the bottom of long page with life story and huge images" is one of the canonical examples of user-hostile web design. EDIT: As in, no one should need a product to tell them this is clearly not what people want.