Last Monday we launched a new product called Mahalo Answers
It got a very warm reception from the press, users and the industry (investors, partners, etc). This was the most seamless and well-executed launch I’ve ever been involved in, so I thought I would share what I’ve learned about launching new products while it’s fresh in my mind.
In this email I discuss coming up with the idea or building the product–that’s almost a book’s worth of information (hint, hint ;-), but I will discuss the moment from when the product is completed, through the beta test, generating buzz, the press tour and launch.
Mahalo Answers is a knowledge exchange with a virtual currency that can be exchanged for real dollars. In other words, it’s a question and answer site like Yahoo! Answers (which was based on the Korean knowledge services by Naver.com or Daum.net), that you can make money from. Users offer a tip that can be rescinded, but rescinded tips are tracked to warn researchers of possible dead beats.
There is no risk to offering a tip as it’s refundable and because multiple users will answer your question and you only have to give a tip to one person. You don’t have to offer a tip,, mind you, but it helps: in the first week questions with a tip were answered in half the amount of time as those without (approximately 30 minutes compared to 64).
For context, this is the third of Mahalo’s five pillars. The first two pillars were our human-selected search results and our Wikipedia-style “Guide Notes.” The other two pillars will launch in 2009. In other words, after almost two years of work we’ve build 60% of what Mahalo will eventually become. So, in three years we will have launched all five products–in some ways companies–in order to reach our goal of building the world’s first “human-powered search engine.”
Now, on to the process.
Creating Buzz: the Beta
In order to create buzz I like to start telling folks about my work schedule about a month out from the launch. On the past couple of trips I started telling folks that I was crushed (true!) trying to finish up “Project A.” At various cocktail parties, meetings and speaking gigs I talked about Project A. When folks asked what it was I told them it was a new product launching on December 15th. We showed some investment bankers and investors, knowing they like to chatter about who’s working on interesting things. Additionally, I started twittering Project A’s impending arrival both in this newsletter and on Twitter.
Many of you played along, responding back to me on Twitter and email and for that, I thank you. 🙂
(editors note: apologies about the formatting from this point)
A couple weeks before the project’s launch we put up a simple Google
Spreadsheet with a bunch of questions. Some of the questions were
important (like what topics do you like), others were red herrings
(like what games do you like to play). In a week about 3,000 signed up
for the beta. We put an NDA at the top of the beta and we started
letting in a handful of people after *personal* emails from me asking
them not to disclose what we were doing and that they were under
official “FrienDA.” People *might* break an NDA if it’s simply text on
a webpage, but I don’t see them doing that if a friend invites them
Additionally, I asked my staff to put their parents and spouses on the
system provided they were NOT in the industry. This is a great way to
get feedback from normal folks, and it’s great for me to get bonus
points with my staff’s parents (“oh… that CEO boss of yours is such
a nice gentleman for letting us see that new product.” — say that
like an overprotective mother for extra effect :-).
In the beta we discussed the type of community we wanted to have and
we put a thin line across the top of the page that instructed users to
simply email feedback at mahalo.com with their feedback. No stupid forms
to fill out–just click and send. I had the feedback email forwarded
to *every* single employee. Why? If there are problems everyone needs
to know about them and get flooded with them over and over again. The
feedback at mahalo.com email is punishment for when we suck–it’s our
penance. Want to get less email? Fix the problems in the beta! 🙂
Side note: I’ve told every member of our team at Mahalo that you MUST,
as a requirement of your job, have a phone with email. If you’re going
to be at a startup you need to be seeing the “Mahalo Weather Report”
on your BlackBerry or iPhone 24×7 like I do. It’s like being a police
officer–you need to carry a piece with you even if you’re off duty in
case sh@#$t goes down. It’s just part of the job.
For the folks who we didn’t get into the beta we sent them a coupon
for five Mahalo dollars to spend in the system. It’s good to do
something nice to the folks who signup but you can’t accommodate if
you do, in fact, have too many beta testers. Many of you took me up on
Slamming the System
It is absolutly essential that you try to break your system before
users come in. I do a lot of simple things like throw bad characters
into web forms or load up 300 web pages at once in Opera and see what
happens. Of course, the tech team can do so much more with the tools
they have. You have to make twice as much time for testing load and
attacks as you think. We got some hacker friends to try and break a
demo copy of the site during the beta, we ran huge attacks against the
system and we maxed out all our infrastructure…
… and still it wasn’t enough!
We had to put another dozen servers online during the launch and we’ve
been up 99% of the time it seems–more than I could ask for for week
The Media Tour (in six acts)
We do 10 media tours a year–at least. These occur under two
circumstances: we are launching a new product or we are doing a
speaking gig in another city. When we were in Sydney, Tokyo, Seoul,
London and Athens this past year we did a blogger dinner and a press
tour. These are invaluable for Mahalo as we’ve actually seen traffic
increases in those countries we visit. If you want to court early
adopters on a global basis do a press tour when you travel. It doubles
the value of your speaking gigs.
[[ If you want to skip to the bottom of the email you can see the
press clippings. ]]
Step One: Selecting who to brief.
Before launching a new product I like to show the product to select
bloggers and journalists. We keep a list in a Google Spreadsheet of
all of our contacts from around the world and when we do a press tour
we look to see which of these folks did a fair job reporting on our
last product. If they really took some time to understand the product
and write an accurate report on it we send them a short note asking
them if they would like to be briefed. We keep it very simple: “We’re
launching a new product and we would love to show it to you over the
phone in the next couple of days. If you’re interested and you could
give me a couple of available times on Wednesday, Thursday or Friday
we’ll set something up.”
Simple. One or two sentences and done. If they don’t respond we might
send one follow-up, but we don’t call and we don’t spam the person
over and over. That looks really desperate and it is not very
gracious. If they’re interested they will let you know, if they don’t
get back to you they are not. Don’t call them over and over to
“confirm they got the email.”
Step Two: Have your hardware and phone perfect.
We set up two computers for every demo. One with Adobe Connect as the
host and one connected as a user. Both computers are right next to
each other on the desk. This allows me, as the person giving the demo,
to actually see what they’re seeing. If the desktop sharing software
is slow or has a problem, I know. I can also pace my discussion to
match what they are seeing. We use a professional headset and phone
system–not a mobile phone. We have the journalists’ phone numbers and
we call them so we know we are getting a great, clean phone line.
Step Three: What to put in your demo.
Before we talk to a journalist we check to see what they’ve written
about Mahalo so we know where we should start the demo. If they’ve
done four stories we don’t need to explain that it’s a human-powered
search engine. This is our job to find out what they’ve covered, not
their job. If they have not covered Mahalo I start by asking folks how
much they know about Mahalo and if they would like me to give them a
brief 10,000-foot view. If they say yes I explain to them: why we
started Mahalo, what we’ve accomplished and what we’ve learned–the
good and bad. I give them specific facts including when we launched,
when we raised capital and how many monthly uniques we have according
to what source (i.e. Quantcast quantified or ComScore’s B.S. numbers
that no one should trust). Give these facts simply and don’t spin
them: fact, fact, fact. Journalists need facts in order to write
stories and in order to form opinions and follow-up questions. It’s
not your job to give them conclusions, it’s your job to give them
After establishing a baseline about the company I explain what we’re
launching in one sentence and why we’re launching it. In the case of
Mahalo Answers I said: “We’re launching a knowledge exchange similar
to Yahoo! Answers, Naver or Daum.net, that has a couple of very unique
features. This is a new feature we’ve added to Mahalo’s search pages
to make them more comprehensive and helpful–we are NOT changing
direction. We’re still a search engine.” At this point the journalist
will probably ask, “how is it different,” at which point you say
“we’ll let’s take a look.”
We then go into the major points that we want to get across. In the
case of Mahalo Answers we had four:
1. Mahalo’s search results contain three items: curated links (like
Google), content (like Wikipedia or About.com), and Q&A (like Naver
Knowledge In or Yahoo! Answers). We believe the future of search is
having these three things on one page.
2. Mahalo Answers has a virtual currency that allows you to “tip” a
user if they give a good answer. Tips are refundable, can be funded
for as little as $1 and users can convert Mahalo Dollars to US
dollars, less a 25% fee. That 25% fee along with advertising is how we
3. Users can ask other users direct questions with a tip or without.
4. Answers can include rich media like a YouTube video, an mp3 file or
a Flickr photo. These are added by simply putting in the URL of the
object–not the embed code.
Step Four: Letting the journalist ask questions.
I leave it up to the journalist as to when they want to ask questions.
After every point I ask “did you have any questions about [[insert
point here]], or would you like me to keep going?” They will ask you a
question or say “keep going.” It’s up to them how they want to ask a
question, not you. Keep moving if they don’t have questions. Some
journalists like to take it all in, or they might be eating a sandwich
at their desk and have you on mute. Either way, give them the option.
Step Five: Wrapping up.
At the end I like to recap the basic points, let them know that we’re
excited about the launch and let them know we have a couple of new
features coming. If they are a smart journalist they will try to
extract one of these new features. If they guess it right I tell them,
if they don’t I don’t tell them. It’s a little game I like to play at
the end which is: if you ask good questions you get rewarded with
better nuggets for your story. At the end I thank them for taking the
time and tell them my email address in case they have any questions or
suggestions at any time. I don’t tell them to talk to a PR person or
contact marketing. If they have a question email me immediately and
I’m available to them. Why? Because when you’re on f@#$%ing deadline
it’s a real pain in the ass to have to talk to a @!#$% PR person to
get a simple question answered. Sorry, I was just remembering my days
as a journalist where I REFUSED to go through PR people and told them
if they wanted INK in Silicon Alley Reporter their CEO needed to talk
to me on email. This made PR people crazy, but it made Silicon Alley
Step Six: Leave them alone.
Do NOT call the journalist to ask them how their story is going or ask
to see the piece in advance. That looks desperate and insulting. If
they take the time to do a piece great, if not at least you know they
are aware of your product for the next time. 90% of the reason folks
hate PR people is because they always act so desperate and act so
annoying. NEVER call a journalist. Period. Just send them short emails
and leave them alone.
Opening the Floodgates
We set an embargo of 1 a.m. PST for Mahalo Answers and we opened up
the product at around 1:30 a.m. We tried to get it to open at 12:55
a.m. but there were–as there always are–some technical issues. We
had a bunch of little problems throughout the night, but had them
worked out when the sun came up.
The entire tech team came in Sunday night as well as half the
editorial team (the other half slept so they could work the next day
obviously). We got a couple of boxes of Stan’s Donuts from Westwood
(peanut butter and chocolate… wha-what?!) and caffeined up for the
night. It was an amazing bonding experience and it was super exciting.
At around 4 a.m. we did a little toast with some sparkling apple
cider, and later that week I packed the team on a big old bus and took
them to Disneyland. One of the gossip blogs took me to task for taking
everyone to Disneyland after having layoffs, but I believe you gotta
work hard and play hard. A day out at Disneyland is a simple reward
for six weeks of non-stop work.
The first day we put three folks on the “sheriff tools” with
instructions to delete any stupid questions or answers, and obviously
spam. We wanted to set a tone in the first week that unhelpful
answers, joking, or obnoxious behavior were NOT the point of Mahalo
Answers. At one point someone sent me (probably one of you!) a long
email trying to figure out why the quality of the answers was so high.
I responded “because we delete the bad ones.”
Membership in an online community is a privilege, not a right. If you
run an online community I suggest removing people who act obnoxious or
stupid–especially early on. Obnoxious folks drive away considered
folks–which is exactly why I left blogging. There simply were too
many obnoxious folks who make their living pissing on the legs of the
folks trying to do something intelligent. Sure, if you piss on a
brilliant person’s leg everyone at the party will notice you, but they
will do so for all the wrong reasons. Anyway, let’s not get into the
de-evolution of the blogosphere or you’ll have to read–and I’ll have
to write–another 3,000 words.
Setting the Tone
Back to setting the tone. It’s absolutely important that when you have
a beta that your entire company take part in it and “eat your own dog
food” as they say. In Mahalo Answers we wanted answers that were more
intelligent than “why don’t you google it?” or “here’s the wikipedia
page.” So, we had our entire team spend days asking interesting
questions and answering them with considerable care and details. When
the users came in for the first time they were confronted with, as our
CTO Mark Jeffrey put it, a library-like environment. It was a serious
place and as such no one started screaming or acting like an idiot. If
they did we deleted their nonsense. Nothing is more effective in
getting rid of a troll than ignoring them and removing their garbage.
You should have no problem removing the bad actors from your system
because one bad actor can cost you, over their lifetime, thousands of
This is how I like to launch a product and it’s not based on anything
I’ve read or been told to do. It’s simply one person’s process learned
from a decade of launching products. I’m sure there are many more
interesting ideas and I would love for you to send them to me so I can
learn from you. That really is the deal that you and I have with this
email experiment we’re doing: I tell you everything I know and I’ve
learned and you hit reply and tell me. We then create a relationship
based on trusting each other, sharing knowledge and support.
I look forward to hearing your tips… just hit reply and you should
see jason at calacanis.com in your “To:” field. I read 100% of emails
sent to me and I respond to at least 50%.
If I don’t speak to before the New York, all the best and let’s kick
some ass in the New year!
jason at calacanis.com