Author here. This is a technique I've been using for a while that I find to be as effective as A/B testing if not more so in moving the needle, trying to gain users for my product stuff.
I'm happy to answer questions about anything I glossed over in the article.
I wonder if you could delve into some of the more long-tail SaaS tracking methods you have found to be successful. Additionally, if there were any pitfalls you found specific to the SaaS space, I would love to know about them.
We are embarking on a new (mostly) discrete segment of our existing SaaS product, and we know that we need to learn more about how users are using our site from the very beginning. Hoping for a big early win, as you said :)
The whole time I'm reading I'm wondering, "Why did he build a relational analytics system? That can be painful to scale." and "Why did he roll his own analytics stack vs using something drop-in like KeenIO?"
The last two paragraphs about signing up for his service instead of rolling your own clarified all that. I'm a big fan of this blog and it is high value, so I don't fault the author for a little self-promotion, but I do wish this would have been mentioned at the top.
Nonetheless it's a solid article and I think rolling one's own analytics system is a fun exercise for the reader, especially when you have time to do so. On the surface it's such a simple schema but making analytics fast at a big scale is challenging.
On Unwaffle - I like that the pricing is based on number of new users per month vs total users. I didn't see the length of the trial period defined on the pricing page (please add). P.S. Trying to load the pricing page in Mobile Safari breaks a little bit — I had to refresh to see it.
One nice thing about SaaS is that scaling problems are almost always a good thing. This was something I actually first built to shake the trial churn out of S3stat, and to be honest there's nothing that would make me happier than to see that trial funnel grow to the point where it starts choking the little analytics ping catcher.
"[personal assistant], be a doll and grab a few hundred thousand out of the petty cash drawer and hire a few contractors to scale the analytics server" (Sent from my BennetonShipToShorePhone)
As to Unwaffle, right now the trial period is "30 days, plus however long it takes Jason to get around to implementing billing". I'll experiment with 60 and 90 day trials once it becomes an issue. For now, though, it's definitely at the beta stage.
Those other things are good general purpose analytics tools of the "throw tons of datapoints at them and they'll give you pretty reports" variety. I think one of them can even track data for specific users.
But they're all things to all people, whereas we're Onboarding & Lifecycle things for SaaS people. So we know and have opinions about things like Trials, paid vs trialing vs expired vs cancelled, etc.
And while you can probably get some good actionable data out of those other tools with a bit of work, it won't be front and center on your dashboard right out of the box.
With our thing, your overview screen will be filled with things you care about, no customization required. And there won't be any extraneous reports lying around that don't fit what you're doing.
9 comments
[ 4.9 ms ] story [ 29.6 ms ] threadI'm happy to answer questions about anything I glossed over in the article.
I wonder if you could delve into some of the more long-tail SaaS tracking methods you have found to be successful. Additionally, if there were any pitfalls you found specific to the SaaS space, I would love to know about them.
We are embarking on a new (mostly) discrete segment of our existing SaaS product, and we know that we need to learn more about how users are using our site from the very beginning. Hoping for a big early win, as you said :)
The last two paragraphs about signing up for his service instead of rolling your own clarified all that. I'm a big fan of this blog and it is high value, so I don't fault the author for a little self-promotion, but I do wish this would have been mentioned at the top.
Nonetheless it's a solid article and I think rolling one's own analytics system is a fun exercise for the reader, especially when you have time to do so. On the surface it's such a simple schema but making analytics fast at a big scale is challenging.
On Unwaffle - I like that the pricing is based on number of new users per month vs total users. I didn't see the length of the trial period defined on the pricing page (please add). P.S. Trying to load the pricing page in Mobile Safari breaks a little bit — I had to refresh to see it.
"[personal assistant], be a doll and grab a few hundred thousand out of the petty cash drawer and hire a few contractors to scale the analytics server" (Sent from my BennetonShipToShorePhone)
As to Unwaffle, right now the trial period is "30 days, plus however long it takes Jason to get around to implementing billing". I'll experiment with 60 and 90 day trials once it becomes an issue. For now, though, it's definitely at the beta stage.
Foxylead.com team :)
Those other things are good general purpose analytics tools of the "throw tons of datapoints at them and they'll give you pretty reports" variety. I think one of them can even track data for specific users.
But they're all things to all people, whereas we're Onboarding & Lifecycle things for SaaS people. So we know and have opinions about things like Trials, paid vs trialing vs expired vs cancelled, etc.
And while you can probably get some good actionable data out of those other tools with a bit of work, it won't be front and center on your dashboard right out of the box.
With our thing, your overview screen will be filled with things you care about, no customization required. And there won't be any extraneous reports lying around that don't fit what you're doing.