What conversion rate validates an idea?
What conversion rate validates an idea?
Much is discussed regarding validation and experimentation, however, one of the key points of an experiment is to validate or invalidate an assumption and thus it requires to establish criteria. A standard experiment is to check how many people sign up after seeing the value proposition on a landing page. Therefore, what conversion rate do you use to qualify your experiment as successful?
I really wish people would stop sharing this myth!
Repeat after me: "conversions on a landing page validate whether you have said the right words to the right people about the right product. Not whether you have a good product."
It's perfectly common to create a good product but get no/very few conversions on your landing page.
Not because your product (idea) is bad, but because you are bad at explaining to the right audience for your product why it will be valuable to them.
Instead, you need to go out and find out who has the problem you're solving and how they perceive and describe the value of your product (and any objections they have).
The most efficient way of doing this is nearly always direct sales.
You can then take what you've learnt and turn it into a semi-decent landing page/marketing campaign.
So, I completely agree with every point you made, but I think it slightly misses the point of the OP (I think?).
You're assuming that the product is built already, in which case, yes, a poor landing page conversion rate is not necessarily a reflection of the quality/need of the product.
However, from OP's post, if ALL you have is an idea, and you ARE finding that you're saying the right words to the right people about the right product for their needs– then that is indeed validation to continue in that direction.
In fact, an even better approach would be to make alternate versions of the page with varying value propositions, split the traffic randomly to each, and see what converts BEST. This can essentially direct your product build from the start.
An alternative takeaway is: don't just trash a whole idea because one iteration of a landing page isn't converting. Try a few copywriting experiments of how you present your value prop. If you can't seem to get anything to perform, you might want to move on to another idea :)
I get where you're coming from.
I'm actually assuming the product is not built yet.
Yes totally agree.
But the probability of each of those factors being true is < 1. Significantly so.
My point is you need to control for your pitch and your audience being wrong, in order to 'isolate' whether the idea is actually right.
So by putting up a landing page and assuming that
low conversion rate != idea is validyou're not actually finding that out at all. All you can say with confidence is that 'at least one of my audience, product idea and pitch are wrong'.If you do direct sales, you can isolate the 'is my product idea valid' variable and get the results you need much faster.
It's plausible that would work in e-commerce (where demand and tangible value are more loosely correlated).
Even then, you need ** a lot** of traffic. 10'000+ site visitors to be able to make a statistically significant comparison most likely. And that's too expensive for most indie hackers.
Again - I honestly understand why people do this. It allows you to avoid direct sales (talking to people, rejection, hard work) and instead create landing pages (fun, creative, feels superficially like a good use of time).
I wish it were, but it's simply not the most efficient way to validate an idea.
When validating an idea you need information to flow to you from the prospective customer. Sales does this. You can ask questions and get detailed answers. Qualitative and quantitative. With followups.
Marketing has the net flow of information from you to the customer. All you get in return are binary answers.
At best, you'll settle into a sub-optimal local maxima relatively quickly.
At worst (and the normal outcome), you'll end up drawing the wrong conclusions after wasting time and money. And you won't even learn a valuable lesson from doing so.
Hope that makes sense!
Validation through direct sales is great if the distribution model is direct sales.
If the distribution model is online ads, and you can't get a landing page to convert, then there is a problem and a great product with a few direct sales isn't going to fix it.
The validation exercise needs to be aligned with the distribution model.
I'm not suggesting that customer input isn't required, but testing a repeatable, scalable distribution model is equally important.
Happy to agree there, it's just that doing this before you have a good product is putting the cart before the horse.
As a founder, you only have bandwidth to do a certain number of things at any one time - and there's no point worrying about scalability/distribution channels when you aren't even sure what product to build and who is going to buy it yet.
I've made that mistake before, and it's pretty much impossible to come back from.
If you've no idea what you're building, then I agree – customer conversations and research are a must before running any tests.
I had assumed the OP meant validating an already formed idea, where the value proposition and distribution has been more solidified.
Agreed - I assumed by 'validate the idea', he meant 'validate the idea I have for a product.'
Great response, and again, I agree on all points! :)
This is probably the top point, as nothing(!) replaces direct interaction with potential customers. Although I think this might be easier said than done for certain consumer products/ideas. But if someone has a B2B SaaS with a high subscription price - most definitely spend your time doing direct sales.
Anyway, talk to your customers/audience, everyone!!
Edit: @louisswiss just saw your post from a few days ago of "Sales > Marketing. Change my mind." I should have read that first, I would have been more prepared! 😄
Lots of words, but no useful information. Whatever you do, you need to define success or failure. Imagine you interview 1000 customers in person, in the queues of several Starbucks around the world. How many of those should be willing to buy your product in order to define your experiment as successful?
The idea of experiment-learn-build is that you can have a rational approach, and that is exactly what is missing in this super long collection of answers that completely missed the point of my question. What is your rationale when establishing what you expect out of an experiment. ANY experiment. Be it a landing page, a prototype, etc. Be it an idea for a startup or a new project within a large corporation.
Maybe this is what's confusing everyone (and I'm finally understanding your original question).. every single experiment has a different rationale/goal, and I'm not going to assign a specific percentage-point constraint to every thing I experiment with.
You just have to do the experiment and see what spits out the other end. And if the math works out where you can be profitable based on those results, you move forward.
Ah ok I see now.
You asked these two questions. I (apparently mistakenly) thought you were trying to ask the same question in different ways :)
Really, I was trying to answer your first question. And the answer to that is "don't try and validate your idea by putting up a landing page and measuring conversions, because there are too many uncontrolled variables at play to be able to make a sound judgement on the validity of the idea"
I'm with you on experimenting and making assumptions beforehand.
What I'd hoped you'd take away from my comment (after all, I'm trying to help you avoid wasting your time and money, not annoy you) is that there isn't much point doing an experiment where you get a result, but have no way of working out why you got that result.
But what do I know :)
Completely misses the point of my question, which was aimed at the general idea of experimentation.
Answered in another subcomment of yours :)
It's interesting, this is obviously a trigger for you :)
But in all seriousness the OP's statement doesn't currently mention, as you've stated:
I'm not seeing the words "good product" mentioned anywhere from the OP. Nor did the OP's original statement ever counter your claim:
While there is nothing incorrect about your statement, the OP's statement isn't a myth either:
That's totally valid, and even aligns perfectly with your statement.
You just did, arguably, a better job of drilling down and more fully articulating it. Doesn't make the OP's a "myth" ;)
Seems a bit harsh...
Thanks for taking the time to develop such an articulated answer ;)
It really is - I hate to see people making the same painful mistakes I did!
Hmm. I assumed
What conversion rate validates an idea?could be taken as validating aproduct idea. A valid product would be a good one.I suppose it's possible to argue semantics - that OP meant 'any idea not just a product idea.'
But that doesn't make sense - how could there possibly be one single conversion rate which validates any generic idea?
This definitely is a standard experiment.
Tying up women and throwing them into rivers to see if they were witches also used to be a standard experiment.
And it was a good experiment - if you wanted to see if women float when immobile.
If, on the other hand, you were experimenting to find out if the woman was a witch, it's a terrible experiment. Because there are too many possible factors at play!
The same holds true for the OP's standard experiment. Great when you're testing one specific thing in isolation against a clear baseline, terrible if you're hoping to ascertain whether your product idea is valid :)
I'm sorry if it came across that way.
As I hope I've made clear, I don't think OP created a myth, I think he just fell for one which most of us fall for (I definitely did) and is curiously stubborn (perhaps because people want to believe it so badly).
So I think what happened is that, like you confirmed, this is trigger for you. No biggie, lot's of things trigger me too :)
What I've noticed is that when I'm triggered I tend to miss the point of the original statement, instead I bring all sorts of baggage from past conversations that were never mentioned and don't apply here.
The notion of a "good product" is so subjective.
My internal definition of a good product can be totally different from yours and everyone else's. And since it was never mentioned in the first place, seems like baggage from somewhere else. Oh-well, no biggie :)
I'm not sure I'm following you on the whole "woman as a witch" experiment, sorry.
But at the end of the day, we do need a way, in a non-academic environment (important point), to gauge interest and improve results.
Does it need to be infinitely precise? No. Does 1 experiment need to produce ALL of the answers? No.
We just need something that gives us more information than we had before so we can move in an educated direction.
If I created a landing page and over 3 iterations of that lander increased anonymous visitors to paid subscribers by 5%, well, that may be good enough to meet my (or my clients) business goals for now. My time may have bigger impact on the businesses bottom line elsewhere at this point.
At least, that's been my approach, and it's worked for me. Not to say it'll work for you or anyone else.
It depends on several factors, not just the %.
You could have a 50% conversion rate, but if it's costing you $500 to get a customer and their lifetime value is $250 then the conversation rate is meaningless.
On the other hand, you might have a 3% conversion rate, with a lifetime value of $250 and it's costing you $100 per customer.
This is the correct response for finding which baseline conversion rate is a success or not, and really the only metrics that are important.
Only reason why this wouldn't apply is if you don't yet have a product to sell, and therefore missing these datapoints.
I am far from an expert on this, but I read Start Small Stay Small by Rob Walling and he mentioned some standard conversion rates:
If your product is 1-50$ you should expect 1-4% conversion.
If your product is 50-1000$ you should expect .5-2% conversion.
The book is from a few years ago now, but I expect that's still more or less the right ballpark.
I'd guess if you're below the minimum then you need to make some changes.
Could you explain the rationale behind this?
They're just some numbers I wrote down from a book I read, they're not mine. He was just giving some ballpark numbers about what to expect.
For me, there is another question to ask with this, and that's how many actions/visitors/etc... are in your experiment?
So if the consensus turns out to be "5% is the conversion rate to qualify as successful" (and I'm just making that number up), the question I need to ask myself is 5% out of how many?
5% out of, let's say, 10 visitors, isn't going to be valid for me.
That said 5% out of 100 unique visitors clicking a button (or whatever) may be statistically significant enough to qualify as a success.
I know when I did a lot of Affiliate Marketing the standard gauge for statistical significance, at the time, was 100.
So if you purchased 100 clicks from Google to your landing page, that would give you a good indicator if you had the right messaging, layout, design, etc... to invest into scaling your campaign.
Assuming the amount is statistically significant... Still, the idea of experimenting is that you should know what output validates/invalidates your assumption, not that you make up the criterion a posteriori.
Sorry, I'm not following your comment at all. I never made a suggestion to make up anything afterwards, nor did I assert any number that is universally recognized as statistically significant. #confused
The goal should be to at least break even on the user acquisition costs. If you spend $10 in ads to get 1 user that makes you $20 you validated an idea. If you spend $100 to get 1 user and make $0 with him then thats not really a business I would pursue further.
This is not the point, given the fact that you don't know the final cost of your product yet. In any case, the idea is how to establish criteria to validate/invalidate your assumptions.
You need to have some idea what the cost is going to be.
Validating a % is meaningless. Presumably you want to validate whether you have a PROFITABLE idea, which means validating if people will buy at your price point, and if you can acquire customers at a lower price than they are worth.
In a book "Start Small, Stay Small" by @robwalling he says that: "typical conversion rates are between 0.5% and 4% depending on your price range and customer base."
So, the idea is that if you get below 0.5% conversion rate then you invalidate your idea?
I don't remember everything, you can read the book, I think for small business its a lot better that hype books like "the lean startup" and etc.
I believe that any product can be sold as long as you package it nicely and convince the users they need it. If you can think of something you need, there are chances that somewhere out there is a group of people that also needs it, they might not know they need it, they might have specific requirements (and its up to you to find those requirements), they might be hard to reach, but nevertheless I think everything can be sold.
So the numbers are just typical conversion rates. I don't think you can base the success or failure of an idea by the conversion rate, but rather the success or failure of the particular implementation / presentation / reach / demographics of YOUR product.
I imagine it would partially depend on your business model. But, yea, good question.
Perhaps you start with a baseline measurement and if it's not where you want it, you attempt to move the needle with new messaging, designs, etc. So all results would be relative to the initial baseline.
My question is exactly how you define the baseline.
The baseline is just where you're at right now. Usually, you want to improve from that position. For me, 0.9 visitors to Brooklet.app sign up for early access and 0.3 of them actually try logging events. That's my baseline. That doesn't actually validate an idea.