Bronte Media

Recently, I have fallen in love with the concept of Customer Development (authority: Steve Blank), which then led me to the derivative concepts of Lean Startups (authority: Eric Ries) and finally the Minimum Viable Product (authority: Venturehacks/Eric Ries). I wanted to put them into practice for what I am doing at Homethinking and here is my first example.

What is customer development?

I’ve written about it here but basically it is the acknowledgment that startups have very little idea if their ideas are good or not. And that’s mainly because startups are trying to solve unknown problems with an unknown solution.

To mediate that fact, startups should launch early and often and measure things that matter to determine success or failure before they jump too far down the rabbit hole or alternatively don’t get excited enough about what they are doing and scale up appropriately. That’s where the minimum viable product comes in. The concept is basically to launch something bare as quickly as possible to help validate customers. That is, whether customers are interested in what you are doing and if you can make money from them being interested.

Homethinking

Which brings us to the problem at hand. I run a site called Homethinking.com. In essence it allows consumer to search for a Realtor to help them buy or sell a home with. We rank the agents based upon their past sales history and customer reviews.

Over the past two years the site has attracted around 150,000-250,000 unique visitors per month. Although that is nice, for the site to reach scale it really needs to do around 1m+ unique visitors. If that happens, life will be all fine and dandy.

So far I have been unable to break out from that traffic range by launching alternative real estate products like our neighborhoods and mortgage sections. I believe I have failed to acknowledge one simple thing in the world of online real estate: All the traffic is in listings.

What I don’t have is $90 million or $30 million in venture capital. Nor a publicly listed entity with $200m in cash and the backing of the national body.

Listings Syndication

So how to compete in such a market? Listings syndication. Because of the 2006 renaissance in online real estate, a shit load of sites were launched with a similar business model: free listings with a premium placement upgrade. That means there is a lot of places to post free listings if you are a Realtor. Listing syndication allows a Realtor to post the listing once and have that listing posted on all the different sites automatically.

A great idea. And one that has been done before. There are three main companies, Point2 Agent, Postlets.com and Vflyer.com. You can see they each get a decent amount of traffic:

So why should I bother trying to compete? Well the traffic justifies the opportunity (calculations a little later) and I believe there exists an opportunity to focus on being a ‘Google Analytics for Realtors’. Roughly here is my vision:

1. Listings syndication (crowded, done before and table stakes)
2. ‘Google Analytics for Realtors’ (surprisingly only Point2 Agent does a decent job. Opportunity to be aggressive and free)
3. Be the ‘Conversion Culture’ of real estate.

The third point involves a complete focus on conversion and bringing the discipline of other categories of online direct marketing to the real estate industry. Here I am imagining a Realtor can access premium property landing pages that are made premium by virtue of their conversion rates. Realtors can dig down into segmenting the listings sites (Trulia vs Zillow etc.). And we can optimize the collective knowledge of listings in general (You should put a picture of the Living Room first vs Exterior shot; Titles containing the words X and Y yield a higher click through etc.)

But that’s all a pipe dream without getting the first two vision points right.

Minimum Viable Product

The first task was to establish that we were able to get Realtors interested in syndicating their listings through our site. On Homethinking, a Realtor can claim their profile for free and edit anything on it. One of the things they can edit is their past sales history. Their current listings are also available to be edited but play a less of a role.

The first thing we did was put a radio button on the edit current listing page. For a Realtor to get here, they had to have logged in and then navigated two levels down for something there has traditionally been little incentive to do.

On the edit page, I put a yes/no ‘Syndicate your listings to Trulia, Zillow and Yahoo?’ radio button on the page. And launched.

In parallel I was emailing with the three and have decent enough contacts at the firms that they were willing to accept a feed but from the Realtor’s point of view whether they clicked yes or no, essentially nothing was happening. A month later, progress has been made with the partners, but our feed is still not live at any of them. This is essentially what the minimum viable product advocates: getting demand validation before launching. It’s a little embarrassing to be honest, but if the opposite was true and no one wanted to tick yes, then it would have provided an early warning system to me.

To my surprise, every person who reached that page in our Realtor Console ticked yes. We were getting around 20 listings a day.

I had proven my first customer validation point. But my strategy of proving the strategy with the three feed partners needed to be revised.

Enter Google Base. Say what you will about the G, they certainly know how to build products. With Google Base, we could create another version of our feed, sign up for a group account and feed the listings through Google Base. The listings appear when folks drill down from the top level search into Maps and Local Search.

The process of creating a feed, validating it automatically through Google Base and having our items live was around 3 days, perfect for our customer development principles.

At the moment we have over 500 current listings and Google Base recognizes the location of around 380 of them. We need to fix that bug and cleanse the user inputed address but it’s a decent enough sample. For the last 4-5 days that we have been live on Google Base, we have received around 20 clicks a day and the listings have been shown about 750 times a day in Google’s SERP for a click through rate of 2.6%.

If we assume 400 real estate listings yields 20 people a day to pages like this, then to get 10,000 people a day with Google Base we would need 200,000 listings. That’s clearly not going to happen in the short term but that is just based on one partner: Google Base. Other studies from listings syndication providers have found that Google Base represents around a fifth of traffic. So that means to get to my 10,000 consumers a day on listings syndication goal, I would need around 40,000 active listings. But even then, maybe not.

The Craigslist Juggernaut

Another customer learning point I have noticed in the month or so that I have started this journey: How much people care about Craigslist. I try to email the Realtor who did tick the yes to listings syndication in their unbeknownst leap of faith and asked them a few questions. And every one mentioned Craigslist. Searching around Activerain on Postlets and Vflyer (two guys that doing listings syndication at the moment and are the market leaders) and there were surprisingly little talk of how Zillow fared against Trulia etc. Instead it was all about getting their listings looking pretty on Craigslist.

I stopped what we were doing and built Craigslist support. Craigslist isn’t like the other feed partners who take in the listing automatically. With Craigslist, a Realtor must go to the site and post the listing themselves (by copy and pasting our html/css into the Craigslist form).

We launched that last week and got our first Craigslist syndication event with Lynn Hermann of Area One Realtors posting a few listings on the Decatur, IL Craigslist. The result? Her 7 listings are getting around 15 clicks a day.

Now, these are all incredibly small numbers but that’s the whole point: validating before scaling up. Learning through user behavior rather than management opinions.

My goals

Even with incredibly small number of days of customer learning with listings syndication, I have a much better sense of the levers, conversion rates and what I need to achieve to get to my goal.

My 10,000 daily visitors to our property pages goal breaks down (with my current assumptions to):

If 400 Google Base Items yields 20 visitors a day, and Google Base accounts for roughly 20% of referrals once all listing feed partners are live, then I need 40,000 active listings. But that doesn’t include Craigslist which looks like it will lower the amount of listings needed if Realtors can make their way through the posting process successfully. So the number is likely closer to 30,000 – 35,000 or even lower.

I can now concentrate my efforts on getting us to that number. In a month I have established that it’s a worthwhile thing for me to build, that Realtors will use it and that we can get some traffic through the partners and distributing our feed.

Thanks for letting me share.