<a very small amount> increased by "311%" is still a <very small amount>, this is a way to oversell numbers, imho.
The rest of the article is not bad, shame about the title.
I definitely agree that the actual conversion rate of traffic to paid user is low -- we're hoping to get that number up to around 1% some day. I wouldn't say it's misleading though since we did actually triple the visit to paid conversion rate.
"we did actually triple the visit to paid conversion rate."
Well...maybe. What are the numbers? If you converted 4 people on 1000, the 95% confidence interval runs from 0.2% to 1%. The difference between 1 conversion in 1000 and 4 in 1000 isn't significant.
But if you converted 40 on 10,000 (vs. 10/10,000), that is significant. So if you've got big numbers to back up your tiny conversion rate, you should just write that into the blog post to silence these sorts of critiques. You don't even need to be specific -- saying that your N is of a certain order of magnitude would be enough to make most reasonable people take you seriously.
He is not saying much about your actual conversion rate. He is saying more that you don't have enough data to make the claim in the title. For example, imagine if your conversion rate was 0. Then one day you got a few sales. If someone made a post claiming 10,000+% increase in ROI, we'd be pretty appalled.
Keep in mind it's a paid, and pricy, service with a very specific type of customer. Don't compare it to conversion rates of other services that have a wider audience. Tripling the conversion rate is huge for them, and no small task.
It would have been more honest to talk about how they increased their conversion rate from N% to M%, this we increased our good rate by Y% or decreased our error rate X% is the oldest trick in the book for academic papers (one that is likely to earn one a well deserved beat down by the PC).
For the sake of a headline, giving a relative difference is perfectly ethical I think. While providing the absolute difference is much more meaningful, it just wouldn't get nearly as much attention. Lying is of course unethical, but a relative difference is hardly a lie, just spin.
Why not do both then? Provide absolute and relative differences, at least let us know what the destination is: we increased our conversion rate by %300 to %.424.
BTW, the article gets this right, its the title of the blog post that leaves out the crucial information.
Netflix isn't really a SaaS because it's not providing "software" (the first "S" in SaaS)-- it's providing content.
On the other hand, Google Calendar, Gmail, etc are providing software and would probably be considered SaaS products. You're providing the content and Google is providing the means of managing it.
Did you know that about 0.3% of the US population has top secret clearance? That's a tiny number. But about 0.01% of the world population have US top secret clearance?
For what it's worth, we did the free premium trial in June. Over 14 days, we received 11 emails from the folks at StatusPage--some about getting started, some (too many) just their regular email blasts, some to say our trial was ending. Eight of these offered an unsubscribe link, which is appreciated, but the default should not be to spam the shit out of potential customers.
I know I've seen other threads on HN and elsewhere about finding the "right" number of emails to inundate your trial users with, so I hope you guys are experimenting with this and tracking what works and what's turning people off.
Still, seems like a solid product that does what it says, and overall we didn't convert because we already had a solution in place that we're comfortable with. Good luck.
Haha yeah we used to send waaaay to many emails. We're now down to 6 over 20 days. Sorry for being overbearing the first time around. When we personally reach out, there isn't an unsubscribe link but as far as I know, all of our automated emails have them.
Email is something we definitely need to experiment with more. It can be an amazing tool when used for good.
As young startups, we all seek engagement and feedback. So we may overreach because we want someone to acknowledge and respond. One way to get fewer emails from a startup -- respond and provide feedback. They really want to improve the service and best serve you. Startups listen and accommodate. Big companies just send canned emails.
I routinely give out "Call the low plan 'Hobby'" advice. Here's why:
a) Price-sensitive customers with side businesses often either affirmatively identify of hobbyists (I would have said that about myself for the first 5 years of my business -- at which point $20 a month would have made it somewhere around the 10th most expensive thing I was buying, by the way) or are willing to buy it without thinking too much about the name.
b) At larger customers, there are social reasons to avoid putting Hobby on a purchasing card / requisition form / etc where one does not have any technical/quota/etc objections against the lowest-tier plan. I've told the following anecdote two dozen times but it's still good: I was once instructed by a boss to get CrazyEgg set up. We were in the $9 a month Hobbyist tier. I put in a requisition for it and he redlined the Hobbyist and price, went to the pricing page, and picked the Enterprise tier (at $500 at the time, IIRC). When I pointed out that I was intimately familiar with the numbers and sure we could make do with Hobbyist, he said (translated loosely from the Japanese) "Eff if I'm going to pass a request for Hobbyist services to [my boss]."
c) See the anecdote above? I can tell you the anecdote. I can't show you the data. But trust me, the anecdote has scaled for companies before.
Actually, I thought about your CrazyEgg anecdote as I was writing my comment. Having worked for a large Japanese company for several years, I did see a cultural difference between what US and Japanese managers were willing to sign off on for purchase req's. But aside from that, I do think a $9/month CrazyEgg plan is within the scope of a true hobby site, whereas $20/month for a status page service isn't. The proof is in testing, but my gut tells me that customers could be segmented without using a plan name that is arguably demeaning.
I've seen businesses choose a plan up from a perfectly good lower run plan simply because the name of it was something like "Solo" and didn't think it fit. It's just clever marketing to anchor people towards the middle.
Unless they actually don't want customers in the lower tier, I think they could channel customers to the middle tier without insulting potential lower tier customers. They could call it the Solo plan, Micro plan, Bootstrapper plan, or even Lil' 'Strapper plan and come off as less insulting.
I doubt they were purposely trying to insult anyone, but with a thousand roadblocks to growth, knocking down one of them by changing a plan name seems like a quick win.
But it's hosted in a different location. So if said business is offline, you won't know why. An offsite location, with backup sites will still stand.
Also, it's not really for people to know... it's for people who depend on it, like API sites. As an app that uses, say Balanced or Stripe, I'd want to know when, why and for how long they are down. So their status is important to me.
Thanks for sharing! At the risk of sounding pedantic, your increase was actually 211%:
(0.424 - 0.136)/0.136 = 2.117 (so 212% actually)
(Increasing something by 100% is doubling it, 200% is tripling it, so on and so forth.)
But to everyone saying he's overselling, if the title had said he tripled his revenue, would you say the same thing? This is a huge accomplishment no matter how you slice it!
Do you guys measure/optimize on "traffic to revenue" or just "traffic to paid"? Does these two measurements have different optimization techniques?
Also, it would be great to learn about how do you generate top of the funnel. As patio11 said there is an "arms race" on that front and I wonder what do you that on that front.
If I understand your question correctly, you're asking if we optimize for traffic -> any plan vs. traffic -> total revenue? Right now it's pretty unsophisticated and we're not really measuring/optimizing on getting people to select a higher plan vs. a lower plan. I'm sure as we nail more of the basics down this is something we'll focus on in the future though.
Generating top of the funnel is actually really hard and something we struggle with a lot. Writing blog posts and posting them on Hacker News (like this one here) result in huge bumps of traffic but it's not really scalable. We added a small "powered by statuspage.io" at the bottom of every status page and are going to do more SEO-minded blog posts.
For early stage startups where there isn't already much of a market (there aren't a lot of companies doing status pages as a service), a lot of traffic and eventual customers come in via word-of-mouth. We have some plans to really give our customers an amazing experience with our product and company as a whole with the hope that it will get people talking about us.
Do a google search for karma based marketing - Colin from Customer.io does an amazing job getting people talking about his company.
Thanks!
One more question:
I like your https://www.statuspage.io/series-t page.
Is there a service which does everything for you or do you guys process this?
We process all of it =/. There are companies that take care of the entire fulfillment process but the volume we're doing is so low that we just mail them ourselves.
We aren't using Google Analytics goal conversion tracking.
Mixpanel was extremely useful just to see where the huge holes in our funnel were. For example, only 30% of the people that see our signup page end up signing up. Now that we know a lot of people are following through with signup, we're focusing attention this week on improving that number.
We just started using Optimizely and so far it's been great but I haven't used it enough to really comment. Optimizely seems to be really great for micro-optimizations and we've been focusing on bigger complete page redesigns thus far.
"For us, there was so much low hanging fruit in watching a bunch of people check out our marketing site. I suspect there's a good amount of diminishing returns there, but if you've never done it, the value you get out of it is tremendous."
I'd strongly suggest that you keep doing actual tests with customers until you actually see diminishing returns. I've regularly seen people think they've fixed all of the big problems - then see another N rounds of "low hanging fruit" appear ;-)
The insights and improvements you get from doing usability testing are very different from the ones that you get through optimisation via a/b testing. Each can inform the other.
Keep doing both until you find you get no value. I suspect that you'll find that you keep getting significantly useful results if you keep it up.
One of the best things that worked for me when increasing conversion rates is to remove all the distractions that's on the page.
To do this, you first need to define what that specific page's goal is (whether it's to have the visitors purchase, share, sign-up, etc...). It'll work remarkably better if you only have 1 goal per page.
From there, you need to look for the things that will distract your visitors from doing that specific goal. It could be the images, an even better content, even the social sharing buttons, etc...
The demos of your app on your site are pretty slick. What are those ? Simple Gifs or is there some other fancy sorcery going on in there ?
PS: i would really like a way to control the playback of these demos. I would mostly arrive when they had already started and frustrated would move on.
The thing that shows off the different parts of the actual status pages is just a big screenshot with explanations controlled with some JS/CSS. The Creating and Managing Your Page section is actually HTML5 controlled videos.
I couldn't figure out why pricing was dependent on number of users. Who are the users? How do I know how many I will have? Do I have any control over who they are?
But most of all, why is this thing tied to that at all?
So statuspage.io displays of a another site is up, down, etc?
I don't get it. If you want something outside your company to do this, get regular webhosting plan from an ISP.
I can see the benefit. It is a totally different architecture to your service. Meaning no chance that the thing that brings down your service is shared by you services servers and the server running your status page.
