Open Source Commentary from Navica's CEO,
Bernard Golden
August 2007
In This Issue
-
Commercial Open Source: Selling Across
the Second Chasm
-
Follow up to last month's newsletter:
The Washington Post: Illustrating the Power of the New
Economics of IT
Commercial Open Source: Selling Across the Second Chasm
Last month’s newsletter
focused on the paradox of the second chasm (See Figure 1).
Playing off of Geoffrey Moore’s archetypal technology
market description, which posits a market chasm between early
adopters and mainstream customers, the newsletter noted a
paradox in terms of the description’s applicability
to open source software.
Figure 1
Moore’s theory states that customers that lie on the
far side of the chasm demand a qualitatively different type
of product from early adopters. These less adventurous customers
insist that the product itself be accompanied by goods and
services that make it easier to put to use: training, high-quality
support, professional services, and so on. In terms of market
breakdown, the less adventurous group accounts for perhaps
85% of the market.
There’s just one problem with this picture as it applies
to open source. Despite the efforts of professional open source
companies to supply those goods and services that are a prerequisite
to mainstream customer purchase, most of them aren’t
actually paying for open source products: the average download
to payment ratio is around 1000:1. I rather cleverly (to my
mind, at least) characterized this paradox as open source’s
second chasm: the chasm of payment. The newsletter explored
a number of reasons for why mainstream companies might be
adopting open source products while not actually, ahem, getting
around to paying for them. (Moore, by the way, has cleverly
sidestepped this exception to his archetype and moved on to
forecasting that IT organizations will soon be separating
application sheep from goats using his theory of core vs.
context – probably more fun than trying to describe
why facts refuse to conform to his earlier theory).
I received quite a bit of feedback about the newsletter.
One person wrote to indicate that commercial open source companies
are all aware of the issue of how to identify likely paying
customers – just as I am aware that we seem to be stuck
in an intractable morass in Iraq – however, being aware
of a problem is not the same as solving it.
Another (from a commercial open source company) wrote to
say that, in his view, there really isn’t a free rider
problem, in that many of the non-customers still contribute
by identifying bugs, participating in forums, or even just
acknowledging that they are using the product, which might
tip another organization over the edge into being a customer.
I completely agree with this perspective, and, unlike many
others involved in open source, believe the community has
tremendous value even if it doesn’t send in checks or
code. On the other hand, I know that commercial open source
companies (even the one that he works at) are methodically
attempting to figure out ways to segment the prospect base
in order to more efficiently focus their sales efforts.
Yet another person wrote in to note that commercial open
source companies undermine their commercial prospects by providing
too-high quality support in product forums. I must say high-quality
community support has never crossed my mind as a disincentive
for purchase. In fact, my sense is that most commercial open
source customers purchase for reasons other than technical
assistance, though, having purchased, they very well might
take advantage of it. Furthermore, I think it likely that
potential purchasers look to the quality of forums to assess
whether an open source product is worthy of considering for
adoption; for the commercial sponsor of an open source product
to avoid forum participation for fear of putting off potential
purchasers might fall into the “cutting off one’s
nose to spite one’s face” category.
I believe the challenge still remains: given a large pool
of adopters, how can you differentiate between potential purchasers
and non-paying (through still extremely valuable) non-purchasers?
And, more crucially, how can you more easily identify potential
purchasers in order to focus sales activities on them?
Here are the things that companies are currently doing to
make their hunting more fruitful:
Know them by coercion
Currently most open source is available for anonymous download;
however, some companies have begun to require registration
for download, which, strictly speaking, does not fall afoul
of any formal open source rules. Frankly, I have mixed feelings
about this. It feels … coercive. More to the point,
it will undoubtedly dissuade some people from downloading
the product and then going on to become members of the community.
For a company like MySQL (not that they require registration,
but using them as an example), losing some percentage of downloads
would not be a big deal, but for early stage companies, in
which the major challenge is getting people aware of the company
and its products and then building a community, losing downloads
by forcing registration seems short-sighted. The notion is
that some of the people registering will be from potential
customers and you can then begin the traditional pipeline
activities to qualify and pursue through to final sale. On
the other hand, attempting to sell indiscriminately may be
extremely counterproductive.
James Dixon of Pentaho has authored a paper
which uses the metaphor of a beehive for open source (an excellent
paper, by the way, I recommend reading it to grasp the totality
of his message); the bees are the community, while people
who want honey are customers. Confusing bees with honey-desirers
annoys the bees; worse, it can drive them off, thereby reducing
the amount of honey available to the customers who want to
buy.
The problem with the coercive approach is exactly what Dixon
outlines: by imposing intrusive burdens on everyone in an
attempt to identify the few, it may reduce trial and adoption
by valuable community members (i.e., the bees in Dixon's formulation),
thereby making the whole product (in Moore's term) less valuable
to potential customers.
Know them by incentive
This is the most common strategy today. The download page
offers the opportunity to sign up to be notified of product
updates, or subscribe to a newsletter, etc. By offering value
to them, some people downloading the product to them will
sign up and become a “friend of the family” that
can be targeted for further interaction. This is pretty much
state-of-the-art today. The individuals who have registered
are offered different types of opportunities to learn more
about the company’s products; based on the interests
expressed, follow-up sales activities are targeted as appropriate.
What more could companies do? After all, the state-of-the-art
doesn’t seem all that whiz-bang, does it?
Here are a couple of other things that are on the horizon:
See who your friends know
One open source company noted that it met good prospects
at the MySQL Conference. By definition, attendees were from
companies willing to spend money to gain benefit from the
MySQL product; they might very well be interested in spending
money with another open source company.
This is a pretty promising direction. The activities along
these lines have been fairly rudimentary – webcasts
offered to one company’s registered base featuring a
second company’s products. Obviously the Open Solutions
Alliance is directed toward this end as well, taking the form
of a vendor association pledged to interoperability, but certainly
seen by the participants as an opportunity to target one another’s
customers.
One thing standing in the way of pursuing this more assiduously
is the strong concern for respecting registrant privacy. If
I’ve signed up for the, say, MySQL newsletter, I might
resent getting the SugarCRM newsletter, to the detriment of
both MySQL and SugarCRM. This is an issue that must be addressed
in order for this sales approach to become more common; however,
given the increasing need of commercial open source companies
to scale their sales opportunities, much work on joint promotion
is likely.
I expect to see more of this kind of thing in the future,
since it makes such sense. A fundamental rule of direct marketing
is that someone who has purchased a product similar to yours
is more inclined to purchase yours; that’s why Lands’
End customers get J. Crew catalogs.
Data mine your registrant base
One person I spoke with said that they were doing rudimentary
data mining of their prospect base – if they see six
different individuals from a single company registered, they’ll
begin calling into the company noting the high level of interest
in the product and suggesting a business relationship. Obviously,
seeking insight from your prospect (and customer) interactions
is extremely logical and holds great potential. The controversy
here is to what extent data mining is actually being practiced
today. I’ve heard some companies maintain that they’re
all over it; on the other hand, many of the companies I’ve
worked with are talking about getting going with a data mining
initiative. As to which I believe, the old saying “who
are you going to believe, me or your lying eyes?” springs
to mind. I guess this should be characterized as a nascent
movement, albeit one with tremendous potential.
Here are the blue sky things that I believe we will see in
the future in the commercial open source world:
Let your product tell you who needs
to be a customer
Several open source companies have put a “phone home”
capability into their product. This enables them to distinguish
between downloads that are never installed and downloads that
end up being installed. Clearly, an installed product is much
more likely to produce a real customer. This “phone
home” capability is very much sub rosa, and somewhat
controversial, but is being explored by these companies as
a way of gaining more insight into who is using their products
(let me be clear, I’m speaking about non-customer instances
contacting the company; paying customer products phoning home
is much more accepted, and often sold as a feature).
We will see much more of this in the future and it will probably
become common in open source products. As noted, the “phone
home” capability is a little controversial, but the
value it presents to the sponsoring company is invaluable.
The targeting potential that this presents is too important
for it to be overlooked.
Also possible will be a further evolution of the “phone
home” functionality. Today the information communicated
tends to be general and mundane: product version, length of
time the instance has been running, etc. In the future, much
richer information could (and probably will) be communicated:
what features are being used, how many users are on the system
at average and peak load, size of database, etc. This is likely
to be very controversial; however, as the nature of open source
users changes toward less ideological and more pragmatic,
the practice might very well become more acceptable. Certainly
the information people are willing to provide for free or
discounted access to goods or services indicates that detailed
“phone home” information might become commonplace.
In fact, disabling the “phone home” feature might
become a saleable item, much like the phone company charges
for an unlisted number.
With a more sophisticated understanding of a prospect’s
use profile, the opportunity to better segment the market
and better focus sales activities much stronger. Furthermore,
an understanding of the use profile of a potential customer
alls a more highly targeted sales proposal to be presented,
which raises the probability of a sale to those companies
identified as members of the more attractive market segment.
Use of richer "phone home" capabilities is going
to become much more common. When users complain about forfeiting
privacy, the commercial open source companies will offer privacy
policies that restrict customer information to the open source
company (i.e., will assuage their concern), but underlying
the response will be the attitude "if you don't want
to use free software, feel free to move to a proprietary vendor."
In one way, this might be understood as a new form of bit
coercion, with information being substituted for money.
Price according to risk
While Moore’s bell curve of technology adoption doesn’t
seem that applicable to commercial open source, it does –
undeniably – reflect one truth: as one moves from early
adopters to laggards, the appetite for risk drops. Simply
put, the enthusiasm for technology is inversely proportional
for tolerance for risk.
If you believe that open source is an unstoppable phenomenon
(and if you don’t, why are you a subscriber to this
newsletter?), then open source will gradually diffuse into
organizations with lower risk tolerance, or, to describe it
another way, a higher need to avoid risk; this need might
be due to organizational cultural factors or perhaps a larger
stake at issue (that is, more to lose, because the company
is larger or is very brand-dependent, e.g., Coca-Cola).
Commercial open source companies are, in the final analysis,
selling insurance. Once you remove bit coercion from the equation,
what you’re selling is support, responsiveness, indemnification,
etc, which all sound like risk reduction. One of things that
has always struck me is not how good a job commercial open
source companies do at delivering risk reduction – the
kudos Red Hat receives about the quality of its support attests
to that – but how cheap they price their service. One
example is MySQL’s all-you-can-eat enterprise deal for
$40K, the same price Oracle charges for a 1 CPU license.
Isn’t this pricing model a bit wrong-headed? It prices
according to the needs of the vendor – how much can
I charge for risk reduction per system – but not to
the needs of the customer – how much risk do I have
to protect myself from? A pricing model that examined what
the risk profile of the customer is and charges according
to that would be much more responsive to real customer needs,
not to mention offering commercial open source companies the
opportunity to focus on higher-revenue customers.
Now before you throw your wireless mouse through the screen,
consider the idea more closely. Variable pricing according
to risk levels is a – perhaps the
– foundation of the insurance industry. Perhaps commercial
open source companies just need to re-evaluate what industry
they’re really in – delivering bits or reducing
risk. For sure, there’s more money in the latter.
Navica News
You can hear me speak at these upcoming events:
September 10, 6:30 p.m., Software Development Forum, Marketing
SIG: "Open source as a business strategy: Alliances,
marketing and development in an open world." More information
here.
September 12, 11:00 a.m.:"Succeeding with Open Source:
Implementing a Winning Governance Strategy", webinar
sponsored by Black Duck Software. More information at Black
Duck Software website.
October 16, 1:30 p.m., GOSCON Government Open Source Conference,
Portland, OR:"Creating an Open Source Policy for Your
Organization" workshop. More information here.
December 11, 3rd DoD Open Conference, Tysons Corner, VA:"Creating
an Open Source Policy for Your Organization" session.
More information here.
If you are interested in having me speak at your
organization:
Contact me directly via email.
You might be interested in my blog postings on CIO.com:
Keeping
Operations Up and Running is Just Table Stakes
Gates:
I Hate it when Google Copies My Strategy
Linus
Torvalds: Virtualization Sucks
XenSource
and Citrix Get Hitched: An *Interesting* Marriage
|
