How to Build an Open Source Business


hello I'm Peter Levine a 16-0 partner

I've been working on open source for

many decades and the last 10 years there

really has been a renaissance in open

source this graph shows the bulk of

investments have occurred over the past

10 years there isn't just more open

source investment this investment is

leading to bigger IPOs and larger M&A


it's interesting to note here that my

sequel which was acquired by Sun


in 2008 for a billion dollars I was

there at the time and I remember

thinking this is the highest valuation

for any open source company from now to


full-stop but look what happened over

the past couple of years cloud era

Mule soft elastic github all multi

billion-dollar M&A events or IPOs and

really to punctuate the sort of two ends

of this particular graph if you look at

Red Hat in 1999 Red Hat went public for

3.6 billion dollars and last year was

acquired by IBM for thirty four billion

dollars open source really started

largely as software for enterprise

infrastructure such as databases

operating systems today open source is

being developed in almost every industry

from FinTech to e-commerce to EdTech to

cyber security further indicating that

it's a really robust time to be involved

in the open source community so what's

behind this Renaissance

sources move from fringe to epicenter

because of what I call a virtuous cycle

of technological and business innovation

on the technical side open-source was

and is the best way to create software

because it speeds product feedback

it drives adoption it improves

reliability all these great technical

characteristics really became a core

framework but as important as this

technical innovation that's occurred

with an open source the commercial

innovation is just as important and I

would argue without business innovation

such as pay for support open core and of

course SAS open source as a service

there really wouldn't be an open source

Renaissance economic interests create

the virtuous cycle or flywheel more

business innovation leads to larger

communities spur more tech innovation

and increases incentives for open source

I would expect this innovation to

continue and I'll touch on my views on

what I call open source 3d at the end of

the presentation

open-source is built on three pillars

they are one project community fit to

product market fit and three value

market fit each of these have measurable

objectives and these pillars sort of

unfold over time so project community

fit is really about getting early

critical community mass indicators here

are things like github stars a number of

collaborators the number of poll

requests an open source can start from

many different places it can start from

inside large organizations it can start

from academia it can start from the

brainchild of some great entrepreneur

sitting and thinking about what comes

next but what's most important is that

open source has a project leader to

drive the effort and typically that

project leader becomes the CEO of

companies along those lines as investors

we are strongly biased towards funding

project leaders since they own the

project they know the product roadmap

and can lead the product roadmap and

know the code base inside note the next

is product market fit and this is all

about organic adoption as measured by

the number of downloads the interesting

thing about product market fit is that

this is the precursor to later stage

sales engagement that these early

open-source users ideally will become

top of funnel leads for value-added

product and services and you can think

about these early users as being an

extension of your organization they

actually help to promote through organic

word-of-mouth the popularity of your

project into the community while working

on product market fit it's important to

think about what will delineate your

commercial product so start thinking how

you will deliver value to someone that

wants to pay for it and that gets us

into the third pillar value market fit

ultimately to build a commercial open

source company finding value market fit

is critical while product market fit

often accrues to individual users value

market fit typically centers on

departmental and enterprise buyers and

is measured by revenue I've noted here

some of the value add features that

enterprises often find interesting

now one pitfall to avoid or be aware of

is sometimes open source can be too good

that is the project and product market

fit is fully complete such that the free

product may not need any paid for

component here are value market fit

business models the first is support and

services this is where you give away

software and charge for support Red Hat

is the example and I would say that the

downside or potential area to be aware

of is this model is very hard to scale

it hard to scale operationally and you

know if you go down the path where this

is your model there's a likelihood that

Red Hat may actually become your

competitor because they really have

nailed the scaling challenges that's

what they do

Red Hat has built an entire organization

phenomenal organization on being able to

support and service open-source users

the next model is the open core model

this is where you have differentiated

licensing around a core open source this

is great for on-prem software the thing

to be careful of here is alienation with

open source versus proprietary and

there's always going to be a tension

there like what do you put in open

source and what do you have proprietary

and how those things interlock with one


and then finally in the SAS model this

is open source as a service it's a

complete hosted offering cloud computing

really unlock the full potential of open

source the hosted service allows open

source to have full strategic value to

the customer when a hosted service is

based on open source or based on

proprietary code the user does not

distinguish between the two and so all

the benefits that accrue to open source

can be done with the full strategic

value of that open source being

delivered to the user and I'd say the

challenge here might be that you arouse

big clouds into doing the same once you

hit a certain maturity public clouds and

licensing will come up licensing is a

heavily debated topic and while it's

important I see a lot of companies

spending too much energy in the early

days on licensing I think it's way more

important to understand your competitive

advantage and use licensing as part of a

go to market strategy and not put

licensing up front and then figure out

what your go-to-market and competitive

advantage should be following the

licensing model furthermore I think

we've over rotated on our concerns about

the cloud and one of the questions that

we've asked our team here at Andreessen

Horowitz can you name a single example

of a company that's been fully displaced

by a cloud provider and the answer is no

and so I think that our concerns around

large cloud vendors coming in and just

co-opting an open source project and

putting you out of business

the concern is probably greater than

reality in many cases open source

companies have become great partners of

cloud companies and approaching it that

way will also help to minimize and

mitigate the concern about clouds taking


there's a couple of things that I would

remind you all of with respect to your

competitive advantage that you have when

it comes to your offering versus public

cloud one Enterprise customers don't

want vendor lock-in two they want to buy

from people who have written the code

and three big companies don't have your

specific expertise they have a lot of

expertise in a lot of things but in

general not specific around your

particular project and again this comes

back to the project leader if the

project leader is involved with the

roadmap and the project and you've

written the code with a group of people

in your company no one can do a better

job with that code and that expertise

than the organization who's built it

when I was at Zen source we had all the

people who built the project and knew

the code and all that and then there was

another company who basically used the

open source that we were creating to

offer a very similar product and every

time we'd go into a customer to the

extent that they knew about our

competitor they'd ask us like well your

competitor says they offer the same

thing what's the difference and we would

say to the customer at the buyer were

the folks who wrote the code we're able

to better support you to lead the

roadmap put new features in and that

shut down the discussion right there

they're like yep we get it that's super

important so the same thing today if you

think about the cloud vendors in that

it's a very tangible competitive

advantage on the knowledge of the code

your ability to go add features and the

ecosystem really built around the

founding people who start the project

on the licensing side a little more

detail on this research shows that the

most projects are using permissive

licenses like MIT and Apache it allows a

project to be inclusive and encourage

his contribution for early startup

founders as I've mentioned my

recommendation is to not stress too much

about determining which license to use

know your options focus on the product

and community and engage legal experts

to help you get the right option most

importantly just to reiterate the point

figure out your go-to-market who your

customer is kind of how you want to go

deploy the software what markets you're

actually going after and then fit a

licensing model into that distribution

strategy now that we've understood the

three pillars of open source let's look

at how to build your organization around

them as I mentioned I think of open

source as a developer driven top of

funnel exercise and one of the things

here it's not about the funnel the next

section here is really about building

specific open source related functions

into your organization and how those

functions result in a funnel going from

awareness and interest all the way down

to revenue building a business is about

connecting that open source funnel to a

strong value driven commercial product

so let's look at each stage

the first is awareness and interest and

this is largely about developer and

community management

it starts with developer evangelism this

drives the top of funnel it's paramount

here to inspire the community ownership

and build a network of technical experts

who promote the project via

word-of-mouth community management is a

key function to communicate road map and

set community expectations branding is a

next early kind of awareness certainly

awareness point and one key question

that will come up is should you brand

your business with the same name as your

open source projects companies have

succeeded both ways but it's important

to know the pros and cons of each

separate names reduce confusion prevents

brand dilution and provides licensing

flexibility data bricks and spark is an

example the same name can help with

drafting from project momentum but risks

alienating open source project members

Zen and Zen source are examples they are

of same name and in all these stages of

the funnel measurement is really

important to understand how to move from

one phase to the other but what's

important here is to measure user

registrations and/or downloads as

evidence that the project is actually

being used one of the things that we

learned I remember this from Zen source

is you get what you measure and we

originally put in our KPIs

we are going to measure downloads but

what we learned is about half the

downloads never completed and of those

half that were completed only about half

were ever used by the user and so

measuring downloads turned out to be

interesting but not sufficient in terms

of a metric that could credibly be used

to indicate the user traction and what

was happening and so at that time we

said ok well why don't we build in some

form of user registration and combine

that with downloads to get a better

of what's happening in the funnel

next stage consideration is really all

about maximizing developer love adoption

and value and this is typically

accomplished with product management

I've come to believe that product

management is a critical role in any

software organization particularly open

source because here's where you balance

not only the bridge between engineering

and your go-to-market organization but

you also manage the bridge between your

community and the roadmaps that you

produce typically in an open-source

company is going to be two roadmaps one

for open source for core functionality

and one closed source for paid features

what I encourage folks to do is to put

your open source roadmap in your

proprietary roadmap literally on one

slide and show the interlock between the

two like when open source gets released

what do you snap from that and put into

your regular product how do those two

things work together I'd also suggest

having a framework to clearly delineate

what will be paid and what will be free

for open source CEOs and for those of

you who are the project leads managing

the roadmaps incorporating customer

feedback and communicating about what is

open versus closed can take a majority

of your time but I think it's very

important to really make sure that you

as the project lead and CEO are working

with your teams to facilitate this

particular activity

next is communication it's extra

important to communicate how you will

invest in R&D and what percentage of

resources will be invested in

open-source versus proprietary

applications features or services and

then third building a growth function

with analytics and telemetry when it

comes to the product itself you want it

to be easy to access and download with

great technical documentation and most

importantly built-in analytics to

understand how users are using and

upgrading the product this helps

identify value market fit from product

market fit the success measure here

should be KPIs about product usage that

predict downstream sales opportunities

and it can be complex to pinpoint these

exact KPIs or really what you want to

develop over the course of time at your

organization or the conversion metrics

at each point of this funnel knowing

that if you start with X number of let's

say developers or users at the top of

the funnel how does that turn into

revenue at the bottom in this particular

area of analytics and telemetry I would

encourage you to experiment with sk use

with product bundling unbundling

bundling you know kind of go back and

forth to find the correct line between

free and paid functionality this is

really a never-ending journey for an

open-source company and the funnel and

feedback cycles have to be very tight

the next stage stage 3 is about

evaluation and intent and here this is

largely about outbound marketing and

sales development on outbound marketing

you want to prioritize campaigns focused

on specific market segments let's say

you're going after development managers

focus the campaign around those users

and these can be ascertained based on

well-known patterns from developer

evangelism and top of funnel inbound but

really focus on those particular buyers

as opposed to generalized marketing


next sales development really takes a

customer success approach to be curious

about developer needs and understanding

what your users are doing here be

inclusive ask questions and don't be

overly salesy and then finally leads get

filtered from legion based on what org a

developer represents and you want to see

if the project is tied to the larger

enterprise or if the project is around

an individual the measure of success

here will be sales qualified leads and

that gets turned over to the go to

market and sales organization for

conversion into revenue which is the

final phase here purchase an expansion

in this particular stage sales may

consist of both self-serve and sales

serve models self serve is typically a

bottoms-up approach and sales serve is a

top-down approach I'll show how these

elements work together but the

coordination of self-serve and sales

serve really works to expand the

opportunities within a customer two

primary activities in your purchase and

expansion phase are new logos and then

expansion an upsell of existing accounts

coordinating top of funnel organic

growth and enterprise sales can lead to

a few common failure modes one is your

open source user doesn't lead to a buyer

and this is an example of where product

market fit might be really great but

there's not value market fit to tie into

it so you have a lot of open source

users but there's no buyers number two

your open source project growth falls

behind your enterprise sales the balance

between what's in open source and what's

in proprietary may be out of whack it

may be here where the value market fit

overrides the product market fit and at

some point you know your organization is

seen more as a closed source proprietary

company as

opposed to an open-source company

similar to that your commercial offering

kills your developer street-cred this

again as an example of where you may

just be putting too many features into

your proprietary product and almost

abandoning your open-source ribs if

successful you may start to see a graph

like this and this is a graph that we

saw at github the x-axis is time and the

y-axis is dollars per customer the

orange line self serve is really that

bottoms up selling to individual users

typically when it's self-serve bottoms

up there's no formal sales organization

that's interacting with a user over time

the or product might be more valuable to

the department or enterprise as opposed

to individuals and that's where these

other lines come into play the green one

is an example of inside sales which

typically will sell to departmental

users and the blue line is field sales

selling to enterprise users the green

and blue lines are examples of top-down

selling and again the orange line is

bottoms up when these are all layered up

and you have the right product for each

of these use cases you can start to

really grow revenue with a combination

of bottoms up and top down selling this

graph is an abstract not every company

will have this exact representation for

example you might not have field sales

to the enterprise you might just have

let's say inside sales and maybe a

self-serve model you really need to

tailor this to your organization into

the markets that you're selling into

so where is open-source headed next 3.0

obviously a slew of headlines in the

last three years show that even

proprietary technology companies are

embracing open-source as an example

Airbnb has open source more than 30

projects Google more than 2000 as

software eats the world the open source

is eating software open source is not

just Red Hat elastic data bricks


it's a part of Facebook Airbnb and

Google and if we think of open source

that way the Renaissance is really only

beginning and the market is far bigger

than we have yet created or yet realized

I like to think that the virtuous cycle

of technological and business innovation

will continue on the technology side

things like artificial intelligence open

source data and blockchain are examples

on the business innovation side we've

had ads supported for a while but things

like crypto tokens and data-driven

revenue are ways of monetizing the

building of open source projects these

are just examples but we can start to

imagine that in five or ten years from

now these innovations will continue on

both sides and as a result of these

innovations open source is a key part of

every company thank you