Ask HN: You're in beta. How much $ is 1,000 random users worth?
10 comments
Correct me if I am wrong but I am going to assume a few things:
- A service offers 1000 users for $X
- The service pays users $Y to join site with the potential to win an iPad/other offer as well as $Y.
If that is the case, the value of users is negative to the startup.
Here's why:
- Users are incentivised to join site, to win iPad thats it.
- They are won't pay to join a service (if its a paid service)
They are not incentivised to join and actually use the product, which are incentivised by acquiring them via Word of Mouth, SEO and other forms of advertising channels. Thus the LTV of the customer is $0 - they won't use the site, they'll just signup as it becomes more profitable for them to join 100 sites and do nothing than it is to join 1 and do something.
Similarly, the service itself is incentivised to actually cheat the startup as they will then get $X and a free iPad or something. They could easily do this by using a bot to join the services.
Even if the service won't do that, there will be users on the service who are actually prepared to do that as well using multiple payment emails etc - if they are paid for joining too!
Essentially by paying for this service, the startup has wasted money by paying for 1000 users who are going to do nothing on their service and have given a free iPad (or other prize).
- A service offers 1000 users for $X
- The service pays users $Y to join site with the potential to win an iPad/other offer as well as $Y.
If that is the case, the value of users is negative to the startup.
Here's why:
- Users are incentivised to join site, to win iPad thats it.
- They are won't pay to join a service (if its a paid service)
They are not incentivised to join and actually use the product, which are incentivised by acquiring them via Word of Mouth, SEO and other forms of advertising channels. Thus the LTV of the customer is $0 - they won't use the site, they'll just signup as it becomes more profitable for them to join 100 sites and do nothing than it is to join 1 and do something.
Similarly, the service itself is incentivised to actually cheat the startup as they will then get $X and a free iPad or something. They could easily do this by using a bot to join the services.
Even if the service won't do that, there will be users on the service who are actually prepared to do that as well using multiple payment emails etc - if they are paid for joining too!
Essentially by paying for this service, the startup has wasted money by paying for 1000 users who are going to do nothing on their service and have given a free iPad (or other prize).
My guess - not much. There are a lot of factors that go into how much a group of users is worth. Do they add a lot of value to your product? I would pay a lot for Woz, Bill Gates, Obama and The Pope to be active users of my product, because their membership would raise recognition of my brand. But just four random people? Worth a lot less. Four random people who are signing up and will forget about my service the next day? Even less than that.
An IPad costs about $500. If you gave one away for every 1000 people, you'd be paying $2 per person to try out your product. Assuming that your product is somewhat niche, you might have a signup rate of 1% (random guess) from a pool of the general population. This would make your cost per user $200.
It would probably be a better idea to imagine who your users will be and target this subgroup specifically. As an analogy, if I'm selling calcium pills to everyone I see on the street, I probably won't sell to many - no matter how good they are. But if I go find people who are afraid of osteoporosis, I will probably be able to sell to many of them.
I would suggest that you stop thinking in terms of absolute numbers of random people (especially if they're not paying) and think more about the value that they can bring to you (whether it is word of mouth marketing, content creation, advice, money or something else). For many products, random people are not the best choice for an initial user base.
An IPad costs about $500. If you gave one away for every 1000 people, you'd be paying $2 per person to try out your product. Assuming that your product is somewhat niche, you might have a signup rate of 1% (random guess) from a pool of the general population. This would make your cost per user $200.
It would probably be a better idea to imagine who your users will be and target this subgroup specifically. As an analogy, if I'm selling calcium pills to everyone I see on the street, I probably won't sell to many - no matter how good they are. But if I go find people who are afraid of osteoporosis, I will probably be able to sell to many of them.
I would suggest that you stop thinking in terms of absolute numbers of random people (especially if they're not paying) and think more about the value that they can bring to you (whether it is word of mouth marketing, content creation, advice, money or something else). For many products, random people are not the best choice for an initial user base.
There's two services here that need to be separated. Doing this as customer acquisition seems to be what most of the commenters here are focused on and does seem like a poor deal.
On the other hand, 1000 software testers for the price of an iPad are just a couple of quarters each. Plus, here's where their unreliability becomes an advantage. Bob's drunk uncle might be a lousy customer, but the aspects that make him a terrible customer make him a great QA agent. He'll forget his password and create fifty new accounts, then demand that they be merged. He'll post comments in EBCDIC. He'll declare he lives in the state of Oregon in the nation of Uganda. He's never heard of Paypal.
After launch, he's a support nightmare and you'll be happy that he disappears the instant that the service is no longer free. However, during the beta, a thousand drunk uncles will find all the edge cases that you never even thought to unit test. They'll force you to tighten down your UI to the simplest possible level of readability. They'll stress test your support infrastructure. And they'll do it for half a buck.
On the other hand, 1000 software testers for the price of an iPad are just a couple of quarters each. Plus, here's where their unreliability becomes an advantage. Bob's drunk uncle might be a lousy customer, but the aspects that make him a terrible customer make him a great QA agent. He'll forget his password and create fifty new accounts, then demand that they be merged. He'll post comments in EBCDIC. He'll declare he lives in the state of Oregon in the nation of Uganda. He's never heard of Paypal.
After launch, he's a support nightmare and you'll be happy that he disappears the instant that the service is no longer free. However, during the beta, a thousand drunk uncles will find all the edge cases that you never even thought to unit test. They'll force you to tighten down your UI to the simplest possible level of readability. They'll stress test your support infrastructure. And they'll do it for half a buck.
Close to nothing, honestly. If the only reason the users are signing up is to win a free iPad (especially with the odds being only 1 in 1,000) I severely doubt that any of them would convert to paid customers.
Now, if you could send 1,000 paid customers my way in that time, I'd be extremely happy to offer you 65% of the initial revenue generated.
Now, if you could send 1,000 paid customers my way in that time, I'd be extremely happy to offer you 65% of the initial revenue generated.
I'd offer them some kind of premium features and/or discounts later down the product (time)line.
Sorry, I revised the question -- I'm basically asking, what you'd pay on top of the premium. To boil it down, I want to know what you are willing to pay to acquire a random user.
Sorry, wouldn't pay a dime. This is a little akin to buying 1000 twitter followers. Sure they'll give you apparent social proof, but you'll have a hard time selling to them. What you are really asking is how much is an unqualified lead worth. You can buy random traffic all day, but if none of them purchase/upgrade its worthless to you.
You have to ask yourself if you really think that you can convert any of these users. There is a world of difference between targeted and untargeted traffic. You probably wouldn't even get any usable metrics out of the deal.
You are also talking about an investment of 2 bucks per user, and you could almost do better with a basic cpc campaign. You'd end up spending a bit more, but you be much more targeted.
You have to ask yourself if you really think that you can convert any of these users. There is a world of difference between targeted and untargeted traffic. You probably wouldn't even get any usable metrics out of the deal.
You are also talking about an investment of 2 bucks per user, and you could almost do better with a basic cpc campaign. You'd end up spending a bit more, but you be much more targeted.
Maybe this will help you, now or later, : http://www.panalysis.com/resources/customer-acquisition-cost...
If you would get value from people signing up, just testing your services you'd probably be better taking the money you would use to buy the ipad and posting on Mechanical Turk.
You'd then get a good number of people to sign up to use your service, they could even do a screencast of their interactions, how they navigate your services and common places where users get stuck.
Asking and paying for direct feedback will be far more useful than buying users with no interaction requirements and hoping for some meaning to arise from your stats.
Least after the testing is done you can purge your system of the testing accounts and then you know who your genuine users are.
You'd then get a good number of people to sign up to use your service, they could even do a screencast of their interactions, how they navigate your services and common places where users get stuck.
Asking and paying for direct feedback will be far more useful than buying users with no interaction requirements and hoping for some meaning to arise from your stats.
Least after the testing is done you can purge your system of the testing accounts and then you know who your genuine users are.
"maybe to test, maybe as actual customers"
There's a huge difference between the 2. You do know you can talk to some mommy bloggers and hold a giveaway, and give 1 iPad in exchange for a few hundred users...
There's a huge difference between the 2. You do know you can talk to some mommy bloggers and hold a giveaway, and give 1 iPad in exchange for a few hundred users...
A random user who is coming to my site to win something who is probably not even in my demographic. And you want me to pay for this?
Literally nothing, or less. Sounds like a worthless headache.
I'd offer a commission to the producer of the traffic, based on conversion after beta. Anyone can generate traffic, and even signups to a free beta.
Getting users that are qualified AND interested that will convert to paid accounts is very difficult, and if whomever is referring users isn't gaming the system, they should have faith that their referrals will convert to paid accounts.
Getting users that are qualified AND interested that will convert to paid accounts is very difficult, and if whomever is referring users isn't gaming the system, they should have faith that their referrals will convert to paid accounts.
1) You'd have to give them some incentive, like give one of them a free iPad or something.
2) You don't know much about the audience you'd get other than that they all have basic English literacy.
On top of the incentive, what would you be willing to spend to get those 1,000 people?