Hacker News new | past | comments | ask | show | jobs | submit login
Why your Referral System Won't Work (georgesaines.com)
21 points by gsaines on July 22, 2010 | hide | past | favorite | 21 comments



I think that there are a couple of factors you are omitting from your analysis.

First, if you take a look at Dave McClure's AARRR model (and if you haven't, you really should [1]), you realize that Referral is something that you only want to attempt after you've got ridiculously satisfied customers. You can't expect people to spontaneously recommend your product on to others until it is absolutely friction-free, and serving that sweet spot so sweetly that they almost can't help it.

Second: not all niches are created equal. Are folks whose itch is scratched by your particular product significantly more likely to be in touch with others having the same itch? For some niches, yes. For others, no.

Finally, in the absence of a Freemium model, you definitely need to have a well-organized Activation process; some way for people to get engaged with the product prior to a commitment.

[1] http://500hats.typepad.com/500blogs/2007/09/startup-metrics...., for example.


Hey Michael, I've actually not heard of McClure's AARRR model, I'll check that out. I've read read some of McClure's other articles, but not that one!

Your first comment about the referral only working in the sweet spot probably did kill it for us. We still aren't at the sweet spot, and the process of engaging with the product definitely isn't frictionless like it is for dropbox.


My good friend recently made me signup w/dropbox so he can get more space, and I'll probably do the same with my friends.

The main reason why it works is because we're helping each other achieve something; we're making an existing process easier/more convenient. It's a MUTUALLY BENEFICIAL activity (ie, it's easy for us to share files with each other).

Dropbox is a platform product like Facebook (ie, join this site, and we can keep in touch together easier). This strategy would work for all platform products.

As for your startup, you have to ask why would I would recommend someone a product that eases remembering japanese/chinese chars, there's no process after that at all. There's nothing MUTUAL BENEFICIAL here at all. If my friend is using your tool, why do I care? It doesn't help me at all.


Very true, jscore, referring people to Skritter has no network effect at all. We do offer extra subscription time to both the recommending party and the new referral, and we spent a long time trying to figure out the right mix of benefits for both, but I think you're right that without the networking effect there's a lot less interest to recommend people.

As you say, it doesn't help you to recommend anyone to a site where the primary activity is independent learning. Now, we could alter that and some social features, but my best guess is that it's not just "social features," because tacking that stuff onto Skritter probably wouldn't help too much! To get maximum gain, we would have needed Skritter to be more easily spreadable from the get-go. :)


Just an FYI -- your page hung on loading "Waiting for dev.visualwebsiteoptimizer.com..." for at least a good 60-75 seconds.


Hmmm, that's no good. I'll have to check that. Thanks for the tip.


Just to chime in, since I discussed referral systems with these guys and on my blog: mine has similarly been an utter failure, netting twenty trials and no sales. Five bucks in AdWords would work better. YMMobviouslyV


Thanks for the comment patio! Our conversation about referral programs is what actually prompted me to write this. If it had just been us, I would have assumed that we were just doing it wrong, but when someone with a lot more experience and skill tried it with similar results I thought "huh, I wonder if there isn't a pattern there."


I think the analysis here is close but perhaps slightly off.

1. Your company has to have freemium pricing. - This is a factor that affects conversion rate. It's not so much about freemium as it is about friction. I think the idea here is that you have to have a more friction free offering for the referral, specifically the offer that the person being referred sees. If they get offered something a little extra than what they could get on their own it should be more appealing.

2. Your product has to have wide appeal. - I think what you really mean here is that your customers need to have people they can make referrals to. This is obviously a big factor in the number of referrals made. Who are your customers? A businessman learning Japanese on his own might just not know anyone that would use the product other than himself. Students might know lots of other students who might be interested - which are you dealing with?

I also like the idea tomjen3 mentioned about incentives for both parties. This helps encourage people to make the referral, do it through the system , and helps the person being referred feel some form of obligation to the referrer and see that they can get something out of it.

I'm not saying that "you're doing it wrong" or that if you make some changes it will work, just that maybe there are some deeper reasons as to why it won't work in your situation.


These are good points ryanelkins, especially about the distinction between "having wide appeal" and having customers that know other customers in the market. This was definitely a problem we faced early on. Essentially we thought we'd be selling to students and instead we've done most of our business with non-students.

Regarding the rewards, we did offer both the customer and the person being referred free time on the site as a reward, but either we didn't do a good job making that clear (very possible) or it wasn't compelling enough. To be honest, we're still not sure which it was.


I definitely agree on #2, that your product needs wide appeal, but #1 (fremium) seems not to be the case.

You can find plenty of places out there that do a refer-a-friend in exchange for the referrer receiving something in return. My gym (24 hour fitness) does this and I have referred people to it before.

I suspect in this case the problem mostly lies in the product being too niche, not that it costs money.


Good point usaar333, I don't know if I did a good enough job making that distinction clear. In our case the niche product is a big referral buzzkill. My thinking though was that niche/non-niche aside, it seems like referrals would spread more easily (they wouldn't necessarily be more monetizable) if there was a freemium version though. Do you agree or disagree with that line of thought?

For instance, if your gym had a "2 weeks free" and didn't require me to take out my credit card, I think I would be more likely to try it out all things being equal.

Then again, my personal experience is often totally biased. I am trying hard to give up thinking that I know what our customers really want and let testing show the way!


I don't buy the theory that the product needs wide appeal at all (#2).

Gym referrals work because it's mutually beneficial, "hey, come join 24h fitness, so we can workout together". I win, and my friend wins.

It's much easier to go the gym with a partner for motivation, etc. than to go by yourself.

Now, let's take travel, which is also not niche. If I'm lonely planet trying to sell travel books, referral won't work because it's not mutually beneficial for both of us to have the same book, etc.

Good discussion going on. I'm actually brainstorming on a referral strategy for one of my sites (travel based).


Mutual benefit certainly helps. I just wonder if you have to have it.

The people I referred I actually never work out with. It was just a trivial way to get $25 or whatever the bonus was at the time.

I believe Zecco (brokerage) also has a referral system going. Again, using it is beneficial monetarily - I could care less though if my friends use it.

One thing that is probably true: If there is no 'network effect', you probably need a much larger reward to get referrals.


I also think that your main page is a lil visually busy, which if we're contrasting against Dropbox, is a major difference that made me less inclined to proceed with the free tour. Also, at the bottom of the page you introduce the founders, which I think is better left til after you've proved the product worth the attention in the first place.


Yeah, Dropbox has a wicked simple homepage, I love it. I'm actually multivariate testing almost everything on that homepage, so it's possible you got the variant with the animation (would make it a lot more busy).

Substituting out the founders bit is something we've played with, but the previous test didn't get conclusive results one way or the other, so I need to keep trying different things with that space.


Agree, there is no need to put the founders pics on the main page - seems distracting and off putting.

Especially as none of them are Asian, and the whole point of the site is to learn Chinese Japanese characters (it lowers my trust in the site)


I think affiliate marketing that you mentioned can be really undervalued to. Sure the business has a shady underside but on the good side of affiliate marketing your paying for results only and hopefully attracting people who spend most of their time marketing to see if they can sell your product.


Also you forget point 3:

It has to have incentives for _both_ parties. I specifically didn't send my mother an invite to dropbox because I didn't want to seem to do it so I could gain something at her expense.

Dropbox actually does give something to both sides, but if I miss that how many others do you think does too?


I've recommended Dropbox to several people, but have only sent invites to those I think wouldn't mind me getting something in return for the referral (and even to those I've told about the extra space we'd get and that they could of course sign up without my link).


The reason Dropbox works is simple: it's a carrot on a stick. Most invites don't have that and provide no incentive, either for the inviter or the invitee.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: