Why Every Business Should Become A Data Business
Peter Hunt, CEO of elementl, on the dagster master plan; Why every business has to become a data business; How to price information products (yes, that’s likely yours!)
This is the Three Data Point Thursday making your business smarter with data & AI.
Want to share anything with me? Hit me up on Twitter @sbalnojan or Linkedin.
Peter Hunt, CEO of elementl, on the dagster master plan
Why every business has to become a data business
How to price information products (yes, that’s likely yours!)
Let’s dive in!
Why Every Business Should Become A Data Business
Funny story, Roundup, the controversial pesticide, was the birth of a huge data network effect-driven monopoly.
And the reason every other business should become a data business too.
Wait, what? In 1996, Monsanto invented “Roundup ready,” a Roundup-resistant soybean. What does that mean?
Farmers could cover their complete fields in Roundup without harming their soybeans and get a 10x increase in their yield.
So how does that connect to data? RoundUp Ready was made…. by genetic engineering.
That means Monsanto had to collect lots of data on what genes do what in which plants to create their super-seed.
That’s the end of the story? Nope! Monsanto then used the data collected from selling these super-seeds to create… more super-seeds, of course!
(Adoption of the Monsanto soybean and other seeds was exponential and 100%!)
The gist of it: The creation of more genetically modified seeds led to more seeds being created due to the increasing amount of data and skills in that field, and thus to more benefit to farmers, ultimately leading to monopoly structures inside the seed market!
That’s data network effects at their best. The Monsanto product range became more valuable to the farmers because more farmers bought them. More buyers => more value to each of them.
Why should YOU care? Because it’s agriculture, duh! If agriculture gets disrupted and forever turned into a market segment where only data businesses thrive, why do you think your market segment is different?
You need to become a data business because someone inside your industry will take that same step. Someone will create an ever-increasing flywheel of:
Making their product better by infusing it with data & crunching data better
Selling it to consumers, and thus gaining the ability to collect more data & crunch more data
Turn this circle round and round again, increasing the value to consumers, and thus adoption, and thus their market share on every turn
Until the segment displays a monopolistic structure. (without you…)
Peter Hunt On The Dagster Master Plan
The data orchestration space is rapidly evolving, raising money (elementl $33m Series B), new entrants (orchestra), and keeps on changing.
Peter Hunt, CEO of elementl, the company behind the open-source orchestrator dagster, recently shared his Dagster Master Plan.
What you should know:
Elementl wants to accelerate data teams’ adoption of software engineering practices, not just (or not at all?) build a data orchestrator.
They realize the importance of data AND machine learning applications; we hope they are looking more toward the latter.
So what? Will dagster deliver on this promise? We think: maybe.
We love the direction, getting rid of the focus on the data orchestrator, the Jenkins Monster of the data world, and focusing on the practices themselves.
We love some features released for this, like data assets, data, and metadata all over workflow.
We don’t love the laid-out plan: Providing minor fixes and doing some polish work to flatten the learning curve - yeah, right, that’s what’s not going to work.
Overall: We think Elementl has figured out a great direction, disconnected from some core assumptions (false assumption “everyone needs a data orchestrator”), and is thinking in the right direction.
BUT they yet have to show real feelable progress for a significant market segment (ideally, we’d like them to focus on production-level machine learning pipelines - but who listens to us…)
So what do you do? Stay tuned, watch them closely, and keep on test-driving dagster; they might push out a genuinely disruptive way of working in the next 1-2 years.
How To Price Your Product
Pricing sounds like the most boring skill.
Dbt didn’t change their prices for years. Slack didn’t, either.
So what? We’re telling you that Slack & dbt are not good role models!
Pricing is like a magic potion; if you do it well, you’re 2x as strong as before.
Boring economics fact: Price elasticity differs. If you change your price by 5% from $50/month, the change in demand is likely very different from what other products experience.
Why care about boring economics? Because it means you have to know about pricing!
You might miss out on a huge revenue add-on, permanent, … And as Mark Andreessen points out, you’re likely to price too low.
There’s a completely underrated book called “Information Rules” written by a genius economist, who happens to be the Chief Economist of Google now for almost a decade.
But ya know, 300 pages of economic theory isn’t for everyone.
Luckily, we read it all (more than once)! So let’s share the basics of pricing:
Stuff on the internet, the 0s, and 1s, is information
Information usually is produced super expensively.
But copies at near-zero prices.
Service another subscriber on my newsletter? Cost = $0
Onboard a new customer onto astronomer.io? Cost ~ $0
That’s a bizarre cost structure! That means you have to price differently!
Doing “cost-based” pricing doesn’t work. Add 10-20% markup on 0? Eeeh.
You must price your product based on what users are willing to pay.
Here’s the catch: That price is widely different for different users!
So your true challenge is…. Guessing the willingness to pay off your different users.
That’s why GitLab CI Cloud is free for me but $99 for an enterprise user: the same product (almost), but a very different willingness to pay.