At the moment I am wrestling with the idea of releasing very early. The prototype will have about 60% of what I view as the “full” functionality. It will have the core features of the service for the individual but will be missing the community features. The current plan is that we get the full drop of the prototype, go through a test and fix phase and then ask in a bunch of friendly faces to try it and and tell us what they think. Part of this would be a discussion on the planned features for the full system to see what made sense to them and what did not. We would then go for the full scalable development for wider release in Q1 followed hopefully by public release later in Q1.
But I’ve been reading (and re-reading) Brad Burnham’s blog post over at Union Square Ventures entitled “Customer service is the new marketing” and his description of del.icio.us really got me thinking. The key paragraph is:
If customer service is the new marketing, it has some important implications in how you build a web services business. First, it means that you need to get the service out there quickly. Designing a comprehensive feature set and spinning it a couple of times in a small closed beta is not going to work. If the service provides value in its initial rudimentary form, your users will take it and run with it. If you launch it fully formed , at best, you have robbed them of the pleasure of co-creating the service, at worst you have created something that nobody wants.
I know exactly what he is saying but the huge concern I have is that you launch too early and it is just too rough to get traction or for people to see the potential. The ability of better funded (aka funded) competitors to out-execute us also weighs heavily on my mind.
I’ll be honest and admit that I just didn’t “get” del.icio.us for a long long time. I think of myself as an early adopter and click-junkie but I just could not see the value in it despite several attempts. I actually think the trigger for me was GMail and discovering the power of tagging. Overnight I became a folder hater and a tag hugger and then del.icio.us made sense.
I see my del.icio.us experience as a problem. I’m not sure I want to release very early and rely on uber-visionary users to understand what we are doing and where we are going. The trick for us is to figure out at what point we have something that just clicks with a wider set of users when they see it and then build on the needs of that user base.
In support of Brad’s thesis, there is also a lot to be learned from the Riya experience of building what they thought the “community” wanted and discovering their entire model was wrong. My guess is that they relied on a pretty homogenous small sample size from the echo chamber to direct the feature set. No-one I talked to outside of the tech world could ever understand what the benefit of Riya was but I’ve described Like.com to several non-tech people and they raved instantly about the idea.
If you have gone down the ultra-early-release route, do you have any words of wisdom or alternatively if you went the traditional route and it failed, would an earlier release have given you more success?
UPDATE 1: Nic Brisbourne has some interesting thoughts on Brad’s post from the VC perspective.
Whilst I am a strong believer in working with customers to refine and build the service, I don’t think that should be treated as the only way to build the business. Over the past few months, the web 2.0 view of everything-on-a-shoestring has been making less and less sense to me. Why would you rely entirely on a viral word-of-mouth approach successfully growing your business when so many other traditional routes are available? Surely you should be trying to maximise every opportunity to market yourself? That costs money and in our business plan, actually takes up a very large proportion of the use of funds. Old school I know but sometimes old school makes sense.
If we have to launch without external funding then that’s what we’ll do, but I know full well that we stand a far better chance of success with money in the bank and a strong marketing plan.


5 Comments
hiya conor
You will know from the 37signals crowd that they go feature poor, build that cut down specification well and then release. That is kindof what we are doing with dbTwang - entry level Phase One that will be tested for 2 weeks by a couple of people (for bugs rather than features) and then released and monitored.
And as there are no absolute rights and wrongs here the route of asking for experienced feedback is a good one!
keith
Thanks Keith, I keep an eye on the 37signals guys.
One thing we could do is put placeholders in for all the advanced functionality and get people to feedback on whether they think they would get benefit from them. This is in preference to giving people a bunch of features as a list and asking them what they think.
The approach I took was to spec my vision of the full system with tons of questions and open issues on the wiki. I then made a copy of that spec and deleted all of those questions and open issues to create the prototype spec. The attitude was “if in doubt leave it out”.
It’s gonna be a fun couple of months!
Looking forward to seeing dbTwang. Will people have to be registered to make any use of it or are “browsers” allowed?
There will be a public gallery for the best of the guitar shots and a search facility that allows you to search the database without registering.
Registration will bring additional benefits
Placeholders a good idea. And have just read that post - nice interview and good points. Ta for reference to it.
keith
Conor,
Good luck with your struggle / launch. It is never easy. Don’t worry about getting it right just admit when it doesn’t work and change.
Munjal
Munjal, reading your superb series of posts made me realise that you have that great combination of a strong vision but with the ability to admit when you are wrong. I hope I can be as clear headed as we negotiate the first few months.