Passive Income Case Study: Haystack.com

In: Business Ideas By: Brian Armstrong

24 Oct 2009

The 37Signals guys launched a new site this week called Haystack.com. Here is their post explaining the launch.

It’s a simple idea: create a directory of website design companies so clients can quickly find a good one to work with.

The idea is not particularly brilliant and original, but they absolutely nailed the execution on this project. Let’s look at a few reasons why this is going to be (and already is) a big success.

At the end I’ll show you exactly how much money they’re making from this just a few days after launch (it’s a lot).

Haystack.com

1. It’s Passive

The first reason this business makes sense is that you build it once and it runs on it’s own. You aren’t trading hours for dollars working with individual clients. What time you do spent on it after launch is spent ON the business, not IN the business. This is an important distinction and one of the keys to building wealth.

2. Freemium

They also used my favorite business model on the web: freemium. People can try it free, but if you want some additional benefits they ask you to upgrade for $99 per month.

This might be worth it to any web design company who gets at least that much new business from it each month. It’s basically a marketing tool for them.

What’s particularly impressive here is that I think they NAILED the breakdown in benefits between free and paid accounts. They didn’t try some gimmicky upgrades like a “pro” badge or a 30 day free trial. Paid accounts get three things:

six images to show your work instead of one (this reminds me of AutoTrader.com)
listings in search results above free users
a bigger display of your work than free accounts (the image is 4x as big)

That last one (size) I’ve never seen in any other freemium model and I think it works great here since they are displaying creative work.

(Side note: another freemium benefit which works really well is an upgrade based on storage, or amount of usage like number of users/projects/invoices sent/etc. DropBox, Flickr, Evernote, Basecamp, etc all use this with great effect, and it works much better than a 30 day trial or something like that.)

3. Permission Asset: Their Blog

The final reason they nailed this is that they have a lot of clout in this field (website design) and their blog has over 100,000 readers.

It’s hard to overstate the importance of this.

The site launched just a few days ago and you can already see on the homepage they have 72 paying customers with big listings (and hundreds of free accounts). That means the site is already generating over $7,128 per month in profit and it’s just a few days old.

What would it have cost to buy this much interest in advertising dollars in just a few days? $100k? $500k? I have no idea. Maybe it wouldn’t even be possible. People don’t put much faith in ads and $99/month is a lot of money. But they got this “advertising” for FREE just by making one blog post. Of course it took them years to built up that asset (their blog) to get it there. And they didn’t fake it by writing cookie cutter, me too, articles. They wrote really useful stuff from the heart.

The message is clear: if you don’t have a blog, you need to create one today and start building a relationship with people who know and trust you. Communicate as personally and authentically as possible. That way, the next time you are doing something important (a talk, a new business, a book, etc) you have a built in audience.

Conclusion

As someone who has a similar business – after all UniversityTutor.com is just another directory site, I’m providing tutors with a way to get new business instead of web design companies – watching the Haystack.com launch gave me a couple new ideas.

Primarily, it’s made me wonder if I have the right breakdown between free and paid tutors accounts on my site. Currently, I let tutors get the first 3 job requests free, then ask them to upgrade for $10/month. This worked much better than a 30 day free trial but still has a problem: tutors who don’t want to pay at that point get kicked off the site!

Currently there are about 7,000 tutors with live profiles on the site. But there are an additional 5,000 tutors who USED to have a profile on the site but they decided not to pay after 3 job requests and got booted off the site. That means I’d have almost twice as many listings if those tutors hadn’t been kicked off (or more accurately, the number of listings would be growing twice as fast over time). It’s also probably hurts the number of tutor signups you get in the first place if people know there is a ticking time bomb on their account (3 jobs left, 2 jobs left, 1 job left…boom! pay up or your account is gone).

So what is the solution? This made me wonder if there are different benefits I could provide for the paid tutor account that didn’t involve a deadline. How about…

Get listed first in search results (yes, maybe)
List more pictures (no, doesn’t work here)
Bigger profile (no, probably not as important here – photos don’t matter as much in tutoring)
Offer more storage, or bandwidth on the site (no, doesn’t work in this case)
Make the job requests limited some other way, like max 1 per month, but you can do that forever with the free account (maybe…)

I’m not sure, what do you think? Getting listed first in search results sounds like a good one, and I might add that anyway (even if I keep the same business model). But I’m not sure if that alone would be enough to convince people to upgrade.

Making a limited number of job requests per month (every month, with no deadline) is an interesting one. This works well for sites like FaxZero.com, where you can only send a limited number of faxes with the free account but you can do so forever without upgrading. So a “light” user gets hooked on it for free, but if they need to start sending more faxes it’s a natural upgrade. But would this work for tutoring? There is a relatively high turnover in tutoring, so you need to keep getting new students, but it’s not like faxes. Getting just a couple good regular students could be all you need for a while.

What do you think? Have I struck the right balance with UniversityTutor’s business model or is there a way to offer a free account without a deadline and curb the number of tutors getting booted off the site?

Until next time, keep breaking free,
Brian Armstrong

Be Sociable, Share!

26 Responses

Markus

October 24th, 2009 at 7:15 pm

How about a larger map marker, premium placement in search results. Maybe something to do with the online tutoring option?

Brian Armstrong

October 24th, 2009 at 9:16 pm

Interesting Markus. Yeah I was just brainstorming and came up with:

When you upgrade your profile will…

* Appear above free listings in search results
* Include reviews from past students
* Be free of advertising
* Include an option to contact you by phone/chat/IM (optional)

Not sure if these are compelling. Also I have mixed feelings about the 2nd one because it would mean removing some reviews from the site.

I don’t have anything formal setup for online tutoring yet, but the one you mentioned about the map marker could work. Good ideas!

Chuck Cohn

October 26th, 2009 at 5:15 pm

Markus/Brian, I’m a big fan of the premium listings being listed above the rest. Brian, I think I bounced this idea off of you 6-8 months ago. Glad to hear you’re considering it.

Brian Armstrong

October 26th, 2009 at 9:13 pm

Haha, yeah I think you’re right. I have a funny way of not listening to people and then months later coming up with it and thinking it’s my own idea :) Go figure. Thanks Chuck!

JONNY | thelifething.com

October 25th, 2009 at 3:36 am

Hi mate, thanks for the post.

Has got me all fired up and inspired to go and and create some more passive business models. I have one that generates enough to live on and I had got lazy.

Cheers and all the best with the Tutors website.

Brian Armstrong

October 25th, 2009 at 3:43 pm

Looks like a great blog you have Jonny. I subscribed!

John Bardos - JetSetCitizen

October 26th, 2009 at 3:28 am

Haystack is a great idea! However, I think it helps to be someone like 37Signals to pull something like this off.

I am trying something similar for English Teachers but it is slow so far.

Brian Armstrong

October 26th, 2009 at 9:09 pm

Most definitely it helped them to have such clout. We all have to start somewhere :) That’s why both you and I are building up our blogs right? Let me know next time you do a launch, maybe I can help.

College Town Menus (CTM)

October 26th, 2009 at 5:08 pm

Great post Brian. I love these examples of your UT site. I can relate to that more than your BV site, but I enjoy both examples.

I think one of the HARDEST things is after development, you may have the best site — but HOW do you get people on-board when they DONT EVEN KNOW YOU EXIST? Granted yours is easier because I’m sure both students and tutors both do a lot of similar Google Searching on same keywords.

In my case, I could have a very scalable system, but it comes down to HOW do I get the RESTAURANT MANAGERS (of each College Town) to know about my site? Sure I could try direct mail marketing (expensive), or Google AdWords (mixed feelings), but not sure how either of those would play out. I want to do exactly what you said, focus on the system/business, and not so much on the individual entities (granted important, but when focusing on the system, the quality trickles down).

I think you have a good set of benefits for joining, but you hit the nail on the head saying that you have 5500 people who have tried, but didn’t continue. When you say you have 7,000 tutors with “live accounts” – are those paying, or does that also include the “pending phase” accounts? Although it sounds like a big number, you still have a pretty solid turnover ratio, so it proves it’s working.

Your proposed new benefits and Markus’ map upgrade sounds great.

Brian Armstrong

October 26th, 2009 at 9:12 pm

Maybe the restaurant managers will know about the site once they start hearing it from customers? If that’s the case, then maybe getting customers is first, and managers will take care of themselves after that. You might need to pay someone to enter the menus for you in that case, although you may already be doing that.

By the way, the 7,000 tutors aren’t all paying! That would be some serious cash, haha.

CollegeTownMenus (CTM)

October 26th, 2009 at 9:20 pm

Haha I didnt think you had 7,000 monthly subscribers — yes that would be A TON of cash!

I agree with what your saying getting customers first is good. But to get customers you need content, and to get the content, you either need to A) go get the menus yourself, type up, and post online with traffic tracking or B) get restaurant on board (which most would require good amount of traffic/usres) first. A vicious loop, classic Chicken vs Egg scenario.

What came first? The student, or the tutor? Haha.

Brian Armstrong

October 26th, 2009 at 9:24 pm

Yup, keep goin! The first ones are always the hardest. It’s all down hill from here.

Rick

October 27th, 2009 at 6:34 pm

Brian, instead of kicking them off the site after a few free contacts, why don’t you keep their profile active and just re-code the “contact” form so that at this point they have to become a paying member in order to read the contact message. They will receive a message from you indicating that they have received a client contact, but that they now need to upgrade in order to view it.

Brian Armstrong

October 27th, 2009 at 9:18 pm

Hey Rick, that’s worth exploring. My only concern here is two things: (1) if they decide not to pay, the clients who contacted them might get irritated for not getting a response and (2) it could turn into a game where the tutors start writing on their profile “don’t use the contact form below, just email/call me instead at…”.

Match.com had a big problem with this. What do you think?

CollegeTownMenus (CTM)

October 27th, 2009 at 9:46 pm

I like his idea. You could easily have an auto response from UT to the student saying that the tutor is not registered and needs to upgrade to view their message. If I were a student seeking help, I would probably naturally contact multiple tutors anyway, so the chance of them only contacting ONE person is little. Maybe in the auto response email, you could say “Contact more tutors [here].” ?? This way it keeps their profile still public (not active), and after they receive probably a few number of contacts after the 3 limit, they’ll realize that UT is legit and will sign up. Great suggestion.

Rick

October 27th, 2009 at 10:07 pm

Brian,

1) I don’t think this will be a huge problem. It may happen that a tutor will not respond to an inquiry because they don’t want to pay, but if they keep getting inquiries they would be silly not to pay the monthly fee for the business leads. I would imagine the breaking point would be relatively low, like 2 inquiries. So my initial instinct is that this problem would be relatively contained. Also, if the student does not get an initial response from the tutor, I am not entirely sure that they would fault your site and get frustrated and seek out another site. On the contrary, they will probably just try to contact ANOTHER TUTOR, which actually helps your business model since it gives you another chance to sign on another paying tutor. But then again, I suppose if 80% of your tutors are not paying a monthly fee, then it would possibly mean that 80% of your student inquiries could go unanswered…not good.

2) This could definitely be a problem. What about a hybrid between your “free trial” and your upgraded features? So new members still have the first 3 contacts free, but once they begin paying they also receive the upgraded features. I think this would solve the problem you reference above about tutors trying to encourage contacts outside of your website. You could also allow users to bypass the free trial stage if they want the extra bells and whistles right away.

Brian Armstrong

October 28th, 2009 at 3:31 am

Good thoughts Rick. I may try it out that way, add the “first in search results” to my existing business model as an additional feature, and build some of the others. That would almost be sure to help. Then if later I want to test having users upgrade solely based on those features vs. limited number of contacts, it would be easier to find out.

Cool! Thanks for the feedback.

Rick

October 28th, 2009 at 2:39 pm

Hope it works for you!

On another note, I was wondering why you chose the subdomain model rather than subfolders (www.austin.universitytutor.com vs. http://www.universitytutor.com/austin). In terms of url keywords for SEO purposes, both models would seem to fit the bill, whereas subdomains seem much more difficult to build and could run you into trouble if there is not enough unique content within each one. At one point I thought you had mentioned that you were linking back from your subdomains to your main domain, with the hopes of building a large number of inbound links, but from what I can see it seems all of your subdomains never point back to your main domain. If you built the site again do you think you would keep the subdomains or just use subfolders?

Brian Armstrong

October 28th, 2009 at 5:01 pm

Hi Rick,

Keywords in subdomains carry a higher weight. Getting incoming links from a bunch of your own subdomains doesn’t work. It was worth it in this case for the SEO benefit (keywords in subdomains) so I’d probably do it again but you’re right it’s a pain to manage an app with subdomains.

Angel

October 29th, 2009 at 9:05 am

The problem is how to monetize these freemium tutors.

Kicking them out is one way of subtly telling them that “If you won’t pay then we don’t want you to stay”. At least from a business point of view its an effecient way of firing off your customers. You dont know how customers feel about the policy.

The sad fact is tutors are also Customers. It really boils down to knowing your customers and giving value to their lives with your products and services. I’m sure you can work something out. More power!

Brian Armstrong

October 29th, 2009 at 3:16 pm

Thanks Angel. Yep I’ll be doing some more tests to determine the exact right match. Kicking them off doesn’t set a very jovial tone to the whole interaction, but it may convert higher than saying “keep the free listing as long as you’d like, but you CAN upgrade if you want a better listing”.

I think if it was me I tend to prefer website that do the latter, but whether it would be more profitable, hard to say. Like I said I’ll do some tests.

Mar

November 1st, 2009 at 7:06 am

In any case, a sudden price tag to a service is always an unwelcomed one. Think of the many sites which offer free, and where once a user base builds up and a price is tagged for continued usage, they’ll flock away to something free that comes along. You started FeedMailPro so you should be all too familiar with this notion.

Coming back to UT and its $10 pricing, you might want to adopt a certain number, like maybe $7, which is more enticing for a user. Of course, it’s your prerogative to keep it at $10, but numbers do make a substantial psychological difference — like how stuffs are often sold at $19.97 instead of $20, or $97 instead of $100.

Interestingly, if you charged $1/mth (which is super duper affordable and when compared to haystack, your customers’ spending power is in perspective here..), would there be a better conversion of tutors? IMHO a $1 upgrade for better listing is a good deal, but what happens when 7000++ tutors upgrade? Do all get the better listing?

Again, it’s your prerogative and with 7000++ tutors, it’s definitely not an easy decision to come up with.

Brian Armstrong

November 2nd, 2009 at 8:18 pm

Hi Mar,

This isn’t adding a sudden price tag, it’s the opposite, removing the price tag where there used to be one.

I split tested $5/month vs. $9.95/month previously, and $5/month performed worse. This tells me $1/month would be even worse. Keep in mind that when going from $9.95/month to $5/month your conversion rate has to AT LEAST double just to break even. So it’s not simply a matter of getting more subscribers.

See here:
http://www.startbreakingfree.com/1003/results-of-universitytutor-com-price-testing/

Regarding ending the price with a 99 cents or something like, I’m currently split testing $9.95/month vs $10/month. No noticeable difference yet in conversion rate, but I don’t have quite enough data yet.

Mar

November 13th, 2009 at 7:09 am

Pardon my error, and true enough you’ll need an inverse proportionate amount of conversion when the price tag is lowered.

As a cursory remark, the $9.95 vs $10, my purchasing consciousness is leaning towards the former. Even though it’s got more digits it feels that much smaller.

James Chillcott

November 4th, 2009 at 8:35 am

Haystack is a great new launch for sure, but it remains to be seen whether it will in fact generate any money for anyone but 37signals. Keep in mind that with all of us on there, the site becomes little different than a localized Google search, although perhaps more fair to the agencies with poor SEO skills.

As the site ramps up, it hits a maximum visibility ceiling (looking like 1st-3rd rankings for most relevant keywords so far), and then 37signals keeps increasing revenues without our exposure increasing. In fact, as they succeed, our exposure should decrease per visit, since there are more and more of us randomly sorted.

Funny how history repeats itself….this is all so reminiscent of elance.com 10 years ago. Great lead gen at first…then the Indian firms caught on (or were favoured by the Indian owners, depending on who you ask) and the site was flooded with lowball offers and terrible lead generation potential.

If we checked out on that 8 years back, why will this work better?

Brian Armstrong

November 4th, 2009 at 2:41 pm

Very valid piont…it could end up like that. They seem to be focusing more on quality than quantity though, and of course have the local component so Indian firms would be in their own category instead of right next to ones in NY.

It will definitely hit a threshold, but this isn’t necessarily a bad thing. Even if elance is flooded, they still make it easier to find freelancers than it was 8 years ago. Probably the same with haystack, even if it’s flooded, it will still be easier to find web design firms than it was previously. If they get ratings and reviews on there that would help, but I don’t see how they could do it since they can’t verify who actually did business with the firm. We’ll have to wait and see! Anyway, thanks for the thoughtful post….you may just be right.
Brian

Breaking Free

Passive Income Case Study: Haystack.com

26 Responses

Leave A Comment

About this blog

Most Popular Posts

You should follow me on Twitter here.