What is a great conversion rate?

Banned 11 years ago

Originally Posted by Joe Ox

1) 51 subscribers on 1150 unique visitors = 4.5%
is this a good/bad/crap conversion rate for opt-in?

It depends on quite a few things, really. I admit I'd be disappointed with it, on any of my blogs. On the other hand, I've also seen lower.

(You're not talking about a squeeze page per se, of course

).

Originally Posted by Joe Ox

2) And what is a good conversion rate for a sales page instead? I've 4% so far but only a few dozens of viewers.

Again, it depends on what sort of traffic it is, how well targeted it is, whether it's "opted-in traffic", "visiting traffic", or what.

But for many kinds of traffic and many circumstances, 4% sounds very good.

Originally Posted by Joe Ox

3) Is the feedback from 200 visitors enough to decide if my strategy is working or not?

Definitely not.

Originally Posted by Joe Ox

How many visits on my landing or sales page are enough to understand how I'm going? 100 unique visitors? 1,000?

2,000+.

Really.

Especially with initial figures like that. This is about "standard deviations". The closer your "strike-rate" is to 50/50, the smaller the sample-size you can get away with, for it still to be "statistically significant". But typically, for a sales page, you're looking at 2,000+. (Many people say 1,000+, and I disagree with them).

Oh well - look on the bright side: you got my post out of the way early, anyway, before others tell you completely different things.

[ 2 ] Thanks
1 reply

{{ DiscussionBoard.errors[7887601].message }}

Joe Ox

11 years ago

Originally Posted by Alexa Smith

Especially with initial figures like that. This is about "standard deviations". The closer your "strike-rate" is to 50/50, the smaller the sample-size you can get away with, for it still to be "statistically significant". But typically, for a sales page, you're looking at 2,000+. (Many people say 1,000+, and I disagree with them).

Hey Alexa, what do you mean by standard deviations and strike-rate?

Thanks
1 reply

{{ DiscussionBoard.errors[7887987].message }}

Alexa Smith

Banned 11 years ago

Originally Posted by Joe Ox

Hey Alexa, what do you mean by standard deviations and strike-rate?

Sorry, Joe - I was talking only about statistical terms.

Strike-rate = the porportion of events ending "successfully" (however defined - in this case, "buying").

Standard deviation explained here: Standard deviation - Wikipedia, the free encyclopedia but - briefly - it relates to "variability of results and what it might signify".

What you want, in these situations, is a "measurement of confidence", and you effectively work it out backwards, asking yourself "how many instances do I need to measure for the results to have, say, a 95% chance of being 'statistically significant' and only a 5% chance of being totally random and telling me nothing at all?"

If the strike-rate is close to 50/50, you don't need very many, but if it's 2% or 4% or something, you need a huge number, to be confident.

It's easy to see why. If somethng happens 50 times out of 100, then 1 or 2 instances being different makes only a tiny difference to the overall conclusion, whereas if something happens say 3 times out of 100, a difference of 1 or 2 is absolutely huge: the difference between 3/100 and 5/100 is a 60+% difference (5 is 66.6% higher than 3, but the difference between 50 and 52 is very small).

So - if your opt-in rate is measured at 50% (it could be, theoretically, from a pure squeeze page), it doesn't matter whether in the long term it's really going to be 48% or 52%. You can measure 100 outcomes and that's all you need to know - that's near enough. But if "it" happens 3% of the time, now you need to measure thousands of instances, for the figures to be meaningful.

Therefore the number of outcomes you need to measure, for your results to be statistically significant, depends greatly on how far away from 50/50 your strike-rate is. And sales outcomes are never anywhere near 50/50, so you always need to measure a lot, before judging.

I hope I've explained it a little rather than making it more confusing!

[ 2 ] Thanks
1 reply

{{ DiscussionBoard.errors[7888042].message }}

Lucian Lada

11 years ago

Originally Posted by Alexa Smith

If the strike-rate is close to 50/50, you don't need very many, but if it's 2% or 4% or something, you need a huge number, to be confident.

It's easy to see why. If somethng happens 50 times out of 100, then 1 or 2 instances being different makes only a tiny difference to the overall conclusion, whereas if something happens say 3 times out of 100, a difference of 1 or 2 is absolutely huge: the difference between 3/100 and 5/100 is a 60+% difference (5 is 66.6% higher than 3, but the difference between 50 and 52 is very small).

So - if your opt-in rate is measured at 50% (it could be, theoretically, from a pure squeeze page), it doesn't matter whether in the long term it's really going to be 48% or 52%. You can measure 100 outcomes and that's all you need to know - that's near enough. But if "it" happens 3% of the time, now you need to measure thousands of instances, for the figures to be meaningful.

Therefore the number of outcomes you need to measure, for your results to be statistically significant, depends greatly on how far away from 50/50 your strike-rate is. And sales outcomes are never anywhere near 50/50, so you always need to measure a lot, before judging.

I was actually looking for a "for dummies" explanation of this, thank you.

I have found an online calculator which can help finding if the difference is statistically significant, but I forgot what formula it uses (a friend of mine who studies math explained it to me, but I forgot): A/B Split Testing Calculator

Thanks

Signature

Professional autoresponder email writer.
I also write premium quality articles, e-books, etc.

{{ DiscussionBoard.errors[7889290].message }}

Nathan Bumstead

11 years ago

I have a few audio trainings from Vince James (the guy who pretty much invented the male enhancement supplement niche and made $100 million dollars) and he said his number is 2,500.

If he can get the numbers to work on 2,500 people, he'll bet his life on it, use credit, and scale it up with money that isn't even in his bank account yet.

Go big or go home I guess... I could never do that myself.

Thanks

{{ DiscussionBoard.errors[7887692].message }}

AntoineM

11 years ago

Hi,

To ease your pain a little bit, here are my numbers:

From Jan. 1st to March 21,
501 unique visitors and only 3 subscribers on my blog.

From Jan. 27 to March 21 (launch the second site later)
408 uniques visitors and 25 subscribers on my squeeze page

Conversion rate on the sales page over 417 unique visitors:
0% - (not all the sales page visitors came from the squeeze page and vice-versa)

Obviously, I'm doing a lot of things wrong and still learning

The thing I found the most difficult is, due to my really low number, it is very difficult to tweak and split test things.

A.

Thanks

Signature

52techniques.com - a blog about business development.
Ultimatecoldcalls.com - eliminate your fear of cold calling.

{{ DiscussionBoard.errors[7887775].message }}

fin

11 years ago

What traffic method?

Thanks

{{ DiscussionBoard.errors[7888026].message }}

smoh

11 years ago

Alexa brought up some good points. Without having your niche or nature of your business it's really tough to say. Here's a few numbers from different styled projects I've done.

Free service site - required sign up in order to use it. Driven strictly through organic traffic, FB, and YouTube. Had ~30,000 uniques/month. Converted at a whopping 25-30% sampled over 3 months.

Ecommerce website selling a service - price range was between 10-100 bucks for a service. Again, organic traffic + social media. Conversion rate was on average 1.6% over 65k + visitors. The 1.6% are completed sales that were delivered. Lots of CC fraud some cancelled orders and other circumstances.

Resource Blog - Not calling this one an adsense site because of negative connotations that come with that. No sign ups, but adsense conversion was at ~3.74% over 10,337 visitors over the last 30 days. The site isn't set up like other spammy adsense sites, but gets good quality traffic for medium to high competitive keyword phrases.

I think your sample size is a little small to know for sure, but it's looking good so far. Try to get a larger sample size and go from there. Make tweaks as you need etc.