In other news, web analytics software is a very crowded marketplace, the competition is stiff. Best not to enter it unless you have something really new and interesting to offer.
Actually it looks like the market may become a very segmented marketplace with a lot of competitors. This is usually a good thing and its a good place to be for a startup.
And its obviously a very big pain that needs to be solved with lots of customers available.
A worse situation is a winner-take-all market with only a few large companies taking 80% of the market: PC-market (Lenovo,Dell), Social Networking (Facebook,Twitter) etc. Here its much harder to gain a foothold but there are companies that do it eg: AlienWare in the highend PC-Gamer market.
I think you're right, but I also believe there are a lot of different kinds of analytics use cases besides a consumer facing website, and to survive many of these players will probably specialize into those more enterprise-y roles.
Similarly, there is room for an upstart to attack those niches.
I've been thinking about building an analytics package geared towards SaaS. I've found it very difficult to track different types of users and their associated revenue in tradition analytics packages (GA, GetClicky, even MixPanel). I've always envisioned a simple analytics package based entirely on user types. Think visitor, free/demo account, paid account 1st level, paid account 2nd level, etc...
I think this is an important missing functionality in many of the existing solutions and I would definitely pay for something that made it dead simple, would HN agree with me?
I've toyed around with a lot of the Analytics packages this article mentioned. None of them do an absolutely great job in the SaaS market, imo.
It's a shame with the rise in popularity of web apps. As crowded as the Analytics market is, there's definitely room right now for someone to turn things upside down.
I used to be a paid Mixpanel user with several accounts. I liked it but it bothers me that you guys don't break the mold any more than traditional analytics did.
Why can't I just choose any random set of events as a funnel and view the breakdown that way? I want to mix and match events on the fly and see how they perform based on past data. I don't always guess the user path correctly, why not help me?
Des has very aptly compiled a list of the leading analytics tools and we’re honored to be included in this list alongside so many great tools (We use and love Chartio and Mixpanel).
I also definitely agree that it would be rather difficult for a single umbrella tool to cover everything, so segmenting the field into traditional web analytics, event-driven analytics, and social analytics makes a lot of sense. We hope to become the industry leader in the rapidly growing social analytics space, but our vision goes beyond that - like you said, Google Analytics primarily drives Adwords sales, whereas with social analytics, there are so many more actionable things you can do. Crowdbooster hopes to guide its users towards that end with our data-driven insights and actionable recommendations, such as telling you the best times to send out content (and giving you an easy way to schedule that content for delivery).
Hummingbird's premise seems like it kind of misses the ball, it certainly would be cool to see a real-time stream of analytics, but what would you actually do with that feature? Seems kind of gimmicky, but I agree with the author in that it could find a home in some other analytics package.
Definitely a solid list of interesting options in the analytics space, thanks for the share.
According to their tech talk about building it (http://www.slideshare.net/mongosf/real-time-ecommerce-analyt...), the original goal was to collect real time data to use in picking products for their gift page. I also heard that the other driving force behind building it was to have a visual way to show visitors to their office what Gilt's noon traffic spike really looks like/means. The core part is under the MIT license on github (https://github.com/mnutt/hummingbird) so it could easily be picked up by some of the other packages out there.
As the author of Hummingbird I'm happy to see it get exposure, but it doesn't serve the same purpose as most of the other packages on the list. It doesn't tell you where you should optimize your site to get the best conversions.
The purpose of Hummingbird is mostly just emotional--it feels really cool to watch your site go from 2 pageviews / second to 400 pageviews / second.
At Gilt Groupe, this happened daily and on an order of magnitude larger scale.
I've often wondered why someone doesn't combine analytics with more traditional executive dashboarding, ie show me revenue per click, per visitor, etc. There is more to most businesses than can be measured in just a website.
Question for HN - does anyone value real time analytics for anything other than a physical LCD dashboard? Personally, I guess it's nice to have, but I wouldn't pay more for it.
Analytics data is only useful if you intend on reacting to it. The same goes for real time Analytics. Its only useful if you intend on reacting to it in real time.
The usage scenario most people fall back on when I have asked them about it is a blog post or something going viral. They don't want to wait the couple hours that Google Analytics sometimes takes to update. Great, you know something is getting a lot of traction with social media or maybe gathering some links... what do you do with that?
If you don't have a plan for how to react in real time, there's no sense knowing your Analytics real time.
(The other usage case often brought up is knowing when there is an issue with your website. In my opinion, this isn't the job of an Analytics package to detect.)
Going real time massively reduced my support ticket queue because devs stopped asking how long it takes for their reports to update!
Aside from that you can use it to identify problems early on - one of my users shipped a broken version of their game and figured it out very early because their level metrics weren't flowing in.
It's also useful for anything with a short lifespan where the hour(s) you wait for GA wastes xx% of an article/promotion/whatever's opportunity.
I would imagine if someone had a business that had dynamic activities in smaller units (like woot off or something comes to mind) it could be valuable, but you're right, ideally, analytics should be too late for anything that makes or breaks your bottom line. It's for looking back at what happened to project into the future, and you have some time to do that.
It won’t be valuable to everyone, but here are some cases off the top of my head:
- On a popular news site or blog, if a particular entry becomes very popular, you may want to follow up on the story on the same day.
- On a corporate or sales site, you may want to track big stories (traffic, particular referrers) as they happen so you can respond if necessary.
I often wonder about the nature of the value of real-time stats also. It's certainly cool and I guess it's useful for certain events (after you tweet something, publish a post, etc.). After that, maybe it would be a very valuable tool for large content publishers—they could see what content is hot and pull it out to the front page, etc?
i'm not sure startup analytics are so different than larger company analytics (i work for a large ecommerce site doing analytics).
might just be my opinion, sitecatalyst can be overly complicated/difficult to implement but it really does have some unique strengths also. it is also one of the more expensive tools though.
http://piwik.org/
http://www.openwebanalytics.com/
Regardless, GA works in most situations at present and will continue to dominate the market. It is due for some much needed upgrades though.