Would you use an analytics platform that automatically identified UX pain points?
Essentially, you would have a dashboard that would help you understand where your users are getting frustrated. Integration would be as simple as loading a library on your website/app with no further requirements.
Expanding on the idea, you could mark certain events as "important" and track paint points related to completion (or non-completion) of said event. For example, understanding why users are abandoning purchases.
The vision for this is to make analytics more useful and automated so you can spend more time improving your product.
Example Use Case & Outcome: You built a mobile app for book hotel rooms. You added the analytics library and patiently wait for something to show up on your dashboard. After a few days, your dashboard shows you that users have trouble entering their information in the checkout page. You rebuild this part of your app to solve the problem which reduces the number of users who abandon the checkout process and increases your revenue.
Follow up question: Would you pay for this service?
I suggest that you watch some videos about talking to customers.
Instead of asking "do you want a website that automatically does this", I think that you should be asking customers more open-ended questions, like "Do your users ever get frustrated? Do you know how? Do you do user testing?" - then figure out the best way to solve the problem after you understand the problem itself. The problem is way more important than the solution.
Brilliant answer. In my experience this is the most common mistake first time founders make.
I would include the question "why is this a problem for you? Would you say it costs you time/money? How much?"
That makes a lot of sense. I will definitely watch some videos and read more about talking to customers. Thanks for the suggestion!
Sounds good on paper but a number of things have to go right for this to work:
Sounds great in theory, but how do you plan on solving this from a technical perspective?
Using your example, how could you tell the difference between a user having 'trouble' filling out the form for technical or UI/UX issues vs. them just not filling out the form because they don't want to? To provide anything of value, you'd need to know how pages are built, how they are supposed to flow and how they're structured...something a 'dropin solution' definitely couldn't do easily without running tons of javascript.
I guess I don't see how you're example provides anything more useful than A/B testing your pages to see if more/less users flow through a particular point.
Just my 2 cents from a developers point of view.
If it worked? Absolutely.
it's something big companies are already paying for. Definitely lots of value but lots of tech challenges too.
I suggest you have a look at what Decibel Insight is doing.
I wouldn't use it. UX is so subjective to the job to be done and what the user values. I don't know how a tool could automatically understand that.
That's fair. I agree with a majority of UX being very subjective. However, there are areas that affect UX that are more objective, like app freezes or users who are frustrated and repeatedly tapping on something). Those objective events would be easier to detect. Albeit, I am not sure how many people actually experience those things.
Sounds great in theory, but how do you plan on solving this from a technical perspective?
Using your example, how could you tell the difference between a user having 'trouble' filling out the form for technical or UI/UX issues vs. them just not filling out the form because they don't want to? To provide anything of value, you'd need to know how pages are built, how they are supposed to flow and how they're structured...something a 'dropin solution' definitely couldn't do easily without running tons of javascript.
I guess I don't see how you're example provides anything more useful than A/B testing your pages to see if more/less users flow through a particular point.
Just my 2 cents from a developers point of view.
From a technical perspective, a checkout form isn't the greatest example. In the context of exception tracking for crashes, think of this as point of failure tracking for UI.
For example, I use a service called AppSee for a startup that is in beta. Some of the things we have identified from watching people use the app (our users know this happens) are app freezes, users repeatedly tapping on things that don't do anything, a dialog displaying incorrectly, etc. To find these things, we had to watch all kinds of user sessions. With apps that have larger user bases, this is impossible.
I hope that clarifies things a bit. I am aiming to solve a somewhat different outcome than A/B testing.
Yes, this is something that would be super valuable. Some companies are trying to solve it, but none of them are doing a very good job.
Google Analytics's new insights tab is full of stuff that isn't useful or super basic. This may be due to use not putting our revenue data in GA.
Amplitude is trying to make discovery of issues easy, but it's expensive and only as good as all of the events you track.
How are you planning to make discoveries? It's probably a good idea to focus on one area of a site (ie. check out page)
My goal would be to automatically capture user behavior (taps, clicks, window close, pauses, etc.) and use some domain knowledge and machine learning to start identifying pain points.
I really like the idea of starting with a specific area to focus on. Checkout pages are an easy sell in terms of optimizing. Any other ideas for compelling pages?
Gotcha, that makes sense.
I like checkout pages as a place to start because the upside is very clear to businesses. Improve it by X% and make $X in revenue. When I used to work at Ancestry we spent a lot of timing looking at which specific field on the payment form did people fall off, rather than just looking at the page as a whole.
Maybe another page would be product details pages on ecommerce sites, looking for the micro interactions that increase people adding stuff to their cart.
Yes!! How would pricing work, Subscription or one time payment?
I am thinking some sort of subscription model. Either a flat monthly fee based on usage. Another model is like to explore based on the number of pain points we identify, but I imagine it would be hard to stay profitable with that model.
Sounds reasonable!!
Yeah, that sounds interesting. It reminds me a bit of Heap’s value prop. Is that similar to what you’re imagining?
Sounds similar. I haven’t used Heap so I don’t know for sure. It sounds like they are automating more of the setup process?