Home About Eric Topics SourceGear

2006-08-24 14:35:46

How to get people talking about your product

Scene:  A Marketing Meeting in the Conference Room

Mickey the Product Manager has called Daisy and Goofy into a meeting to discuss promotional plans for their upcoming new product.


Mickey:            Welcome, Team!  Time is growing short.  Our product is now in beta, so we need to decide how we're going to do the marketing.  Any ideas?

Daisy:               Maybe we should run a bunch of ads in magazines?

Goofy:              Yeah!  That's a great idea!  The professor in my marketing class told us that magazine ads are one of the most effective forms of promotion.

Mickey:            Actually, I saw a post on Eric Sink's blog which said that all magazines are going to be out of business by sometime this fall.  Any other ideas?

Daisy:               Maybe we should do some word of mouth marketing?

Goofy:              Yeah!  That's a great idea!  The professor in my marketing class told us that word of mouth is one of the most effective forms of promotion.

Mickey:            Excellent!  It's settled then.  We'll use word of mouth marketing to promote our new product.  Great work, Team!

Word-of-mouth is not a strategy

Everybody knows how powerful word-of-mouth and buzz can be.  Lots of ink has been spilled about all the nice things that happen when people are talking about your product.

But this is a little bit like explaining the nice things that rich people can buy.  It is obvious that good results are good.  We don't need to be told that.  What we need to know is how to make those good results happen.

Word-of-mouth is not a strategy.  It is the result of a strategy.

How do we make it happen?

Begging bloggers for links

In the last couple years, the most popular way of trying to generate word-of-mouth marketing is a two-part strategy that looks like this:

  1. Make a list of all the top bloggers.
  2. Send email to them and ask them to mention your product.

This won't work.

In fact, any marketing approach which is a post-processing step is going to give very disappointing results.  Recall that there are two phases of marketing:

If you want to generate buzz, you have to start early.  The foundation of buzz lies in the decisions you make in phase 1, not in the tasks you perform in phase 2.

How to Get No Buzz

The most popular two step product strategy is:

  1. Find a Very Large Group of People
  2. Build a product and hope that 5% of them will buy it.

This won't work.

It seems like it should work.  We chose a Very Large Group of People to make the numbers look good.  Five percent of a big number is a lot more revenue than five percent of a small number, right?  And if our product has enough features, surely one out of every twenty people will like it, right?

The result of this approach is that you don't have enough focus to make good feature decisions.  The Very Large Group of People is simply too diverse.  You can't decide which features to leave out because every time you consider a feature, you can think of at least one person in your Very Large Group of People who might want it.

Your best case result with this strategy is that you end up with a product that everybody likes, but nobody loves.  Nobody is going to see your product and believe it was perfectly designed just for them.  Therefore, nobody is going to call, email or blog their friends and tell them how great your product is.  Bottom line:  No word-of-mouth.  No buzz.

Instead, focus on a smaller group

The better approach is far less intuitive, but far more effective:

  1. Find a Very Small Group of People that share common interests
  2. Build a product that 100% of them will love

Numerically, this sounds wrong:

The reason this approach works is because it is much easier to delight a small homogenous group of people than a large diverse group of people.

And your product will ship sooner because you only need the features that are required by the folks in your Very Small Group of People.

And most importantly, when your product does ship, you will have set the stage for buzz and word-of-mouth to happen.  If your product is so perfectly designed to delight your customers, they will be dying to talk about it.

Build a product that is worth talking about

People don't talk about products because somebody asked them to.  People talk about products because they're interesting.  They talk about products they love.

Recently Scott Hanslemann posted a blog entry about a product called FinePrint.  Why did he do this? 

The first step in getting people to talk about your product is to have a product that is worth talking about.  Design something insanely great.  Think about the folks in your Very Small Group of People, and build a product that they will love so much that they can't keep quiet about it.

An example of great word-of-mouth

Delphi is a programming language and development tool from Borland.  It is very highly regarded by its users.  People often speak of how it allows developers to be extremely productive and how it generates very efficient standalone Win32 executables.

Borland is currently in the process of spinning out its developer tools into a new company which is informally referred to as "DevCo".

Three weeks ago I published chapter 9 of my Source Control HOWTO, in which I rather whimsically exaggerated by saying that Eclipse and Visual Studio are the only two IDEs left in the world.  I figured that this might ruffle the features of anyone who is using one of the many other IDEs that admittedly still exist, but I didn't expect any sort of backlash.

Nick Hodges, product manager for Delphi, responded by asking his readers to go to my blog and post a comment in protest.  As a result, 15 raving Delphi fans posted their complaints about Delphi and Borland's products in general.


Fifteen people?  That's like 20% of all the remaining Delphi users on earth, right Nick?



Truth be told, I was impressed.  I've always heard that Delphi is a great product.  It's clear that Delphi's fans are a great group of people as well.  They're passionate and they're tightly knit.  DevCo has some big challenges ahead, but don't count them out as long as their customers love the product as much as they apparently do.

But I was even more impressed with the Delphi community last week when I saw this post by Tate Needham from FinalBuilder.  Basically, he announced the results of a survey in which he asked his customers "Which Version Control systems do you currently use, or plan to use in the next 12 months?"  The top ten responses were as follows:

  1. Microsoft Visual SourceSafe
  2. Borland StarTeam
  3. Subversion
  4. FreeVCS / JediVCS
  5. CVS
  6. SourceGear Vault
  7. QSC Team Coherence
  8. Microsoft Team System
  9. Perforce
  10. Rational ClearCase

I think this data is very interesting in several ways, but I want to zero in just one issue:  Why is QSC Team Coherence ranked so high?

I mean no disrespect to Team Coherence.  It seems like a fine product.  One of their people, Ewan McNab, is a regular on the same forum where I hang out.

But seriously now -- Does anybody really think that Team Coherence is the seventh most popular version control tool in the world?  How could QSC's product rank so high in this survey?

Here's why:

If a similar survey were conducted with a much larger and more diverse group of developers, the results would almost certainly be very different.  For example, Perforce would rank higher, while Team Coherence would rank lower.

Please note that I am not crying foul or bias or unfairness.  In fact, I am delighted (and quite surprised) that my product ranked as high as it did among FinalBuilder developers, especially since Vault is not open source and has no particular connection to the Delphi community.

Furthermore, I am not intending to disparage any of the products on this list.  All the version control tools on the list above are fine tools and worthy competitors.

Except of course for SourceSafe, which everybody knows is a steaming pile of cow dung:-)

My point here is simply this:  The community of Delphi developers is very strong, and they talk to each other.  They're a perfect example of a Very Small Group of People.  Build a product that they will love, and before long, they won't be able to keep quiet about it.

Postscript:  Preemptive responses to comments

"Eric, how can you say that the Delphi community is a Very Small Group of People?  There are 500,000 of us!"

Neopets has 70 million users.  MySpace has 100 million users.  Visual Studio has 5 million users.  A group of 500,000 people is Very Small.  And yet, it's a big enough market to sell a product into.

"Eric, how can you say that people choose products based on what language the product is written in?  People just want software that works.  They don't care if it's written in C#, Delphi or Erlang."

Normal people think that way.  Many developers don't.  A developer whose primary community affiliation is around a programming language with minority market share will definitely be more inclined to buy a developer tool that is written in that language.

"Eric, you forgot to point out that FreeVCS/JediVCS is written in Delphi as well."

No, I didn't forget.

"Eric, how can you say that we have to try and get 100% of our market niche?  That's impossible, right?"

Right.  The point is not that you have to get absolutely every customer.  The point is that going after all of a niche will help sharpen your decision making.

"Eric, did you really say that all magazines are going to be out of business sometime this fall?"

No, but Mickey seems to think that's what I said.  On the other hand, Mickey is obviously an idiot.

"Eric, have you even used Delphi?"

Nope.  I just think the Delphi community serves nicely as an example of my point.

"Eric, if you think the Delphi community is so cool, why doesn't your product support any integration with Borland tools?"

Er, good question.  Maybe we don't get very many requests for this feature?  And maybe that's because Team Coherence is doing such a fine job serving the needs of Delphi users already?

"Eric, how can you say that all we have to do is build a great product and the word-of-mouth will just happen automatically with no marketing communications effort at all?"

I didn't say that.  I admit that there are things that can be done in the marketing communications phase to help generate buzz.  I'm just saying that being interesting is a necessary condition for buzz, but it is not necessarily a sufficient one.

"Eric, I'm in charge of promoting a product that isn't interesting.  You're basically saying that there is nothing I can do to generate buzz.  You're saying that all of the real mistakes were made a long time ago when the initial product decisions were made and now it's too late to fix it.  How does this help me?"

It doesn't.  Probably, nothing is going to help you.  Don't shoot the messenger.

"Eric, how can you say those terrible, unkind things about SourceSafe?"

Oh, settle down.  I'm just kidding.  SourceSafe works just fine for lots of people.  Its bad reputation is partly deserved and partly exaggerated.  In part, that's because exaggerating is fun.

"Eric, have you even seen a steaming pile of cow dung?"

Yep.  Here in the Midwest you can find them all over the place.