If you are one of those fortunate people that know us personally ;) it all seemed so obvious that we would start something. Jack, Andy and myself are figurative peas in a pod. Our entrepreneurial mindsets, hands-on technical ability, and core moral values are more in sync than any other cofounders I’ve ever had the pleasure of doing business with.
But a few months ago, it was anything but. Jack was embroiled in the archetypical management consultant lifestyle; Andy was an engineer at a well-funded startup; and I was building out Disney’s analytics stack. Our opportunity costs for doing this were high.
So what made it all happen?
Well, it’s hard to know when exactly it really clicked for us. Conceiving of a vision is not a binary event; small concepts and ideas pile up over time. One thing led to another and before we knew it, the Collective became all we could think about.
Jack and I were working on an extended engagement in Sweden, and we were knee deep in work. But we did find the time to catch up, and several sauna chats and nightcaps later, together with Andy we worked through a business canvas and decided to make a bet.
The Data Scientist’s PoV
We first looked at the data scientist point of view. What prevents them from finding work they want? What makes them unhappy, and causes them to look for other jobs?
If you are not a data scientist and you’ve ever hired or worked with a data scientist before, you would know they are a different breed of employee. Let me spell out some of the ways:
- Relentless desire to learn and be engaged with interesting problems
- Perceived lack of loyalty
- Bias for action (“let’s just try it and see”)
Of course, the above doesn’t characterize every data scientist, but my point is that they differ from the traditional employee and that for most organizations, some concessions need to be made to accomodate to their working style.
When you hire a data scientist in-house, they don’t like being stuck on one problem or to maintain software for long periods of time. They tend to like to rotate projects and engage with the business holistically, whether with business developers, database administrators, or CxOs. They tend to dislike deadlines, as that can water down the end product or vision for their model or algorithm.
But in practice, companies often can’t make these concessions. We boil down the data scientist problem as follows:
- The data scientist’s agenda is often different from the organizational agenda. They often don’t like adhering to existing deadlines, project management norms, or cadence of change
- Data scientists can spend up to 90-100% of their day wrangling through data versus actual analysis. They often aren’t effective without a technical leader
The Operator’s PoV
Things from the business operator’s vantage point aren’t any rosier. What are the core pockets of value that a data science consultancy brings to the table? How do we differentiate ourselves from the myriad MLaaS and purported one-stop-shop solutions providers?
Our first observation was that as an industry, data science and machine learning is fairly symmetric. There are plenty of talented people willing to work, and plenty of companies willing to hire.
But what leads to inefficiency in a market with otherwise healthy supply / demand dynamics is an information barrier. Operators are willing to pay exorbitant fees to ablate what they don’t know, and snake-oil consultants are unfortunately taking advantage of it.
Now does something feel wrong with that? Precisely because it could not be any riskier to pay exorbitant amounts to become informed about a topic you don’t know much about. Yes, you are paying for a service, but most of the time, you know exactly what you’re getting when you do that.
It sounds obvious but when you go get an oil change, you have little to no doubt about what happens: that once you drive into the shop, a trained mechanic will open up the hood of your car, replace your oil, and you drive out of the shop happy. No ambiguity is present regarding the problem to solve, process involved to solve it, and the end result.
In data science, however, not only do our clients often not understand the work, they are often betting on its transformative nature, while not really knowing exactly how that can happen. Data science itself is also fundamentally an R&D activity - most use cases are not “plug and play” - they need to be tailored to a business’s specific needs. It’s like driving blind - akin to making an investment in an exotic security where you don’t have a good handle on the risk.
On top of all this, as an operator, you’re taking the inherent personnel risk when you hire a consultant, or build a team - how do you know that these are the right people for the job?
We boil down the operator problem as follows:
- You don’t know what you don’t know; many operators make investments blind
- High demand for data scientists leads to exorbitant fees; higher fees don’t necessarily mean a better service
- Organizational paralysis or inflexibility strongly affects data scientists. If key players leave, they often cannot be replaced
- There is no standardized offering in the industry; all vendors offer slightly different solutions to largely the same problem
The Collective Solution
But what if we could create a two-sided market - one where the best strategists, scientists, and engineers in the world could work on the most interesting problems for the best companies, on their own time? What if despite that, clients could be happier than hiring in-house or with big consulting?
More radically even, what if this wasn’t forced labor or employment, and that we only band together out of common interest, to be able to live an inspired and fulfilling life out of doing the thing we love?
Our bet to make data science consulting work is by creating a virtuous cycle of:
- Interesting, sustainable work that is value creating for our clients;
- A pipeline of the best scientists and engineers; and
- A culture with sufficient time for research, learning / personal development, and rest
Consultants traditionally work long hours, are constantly pushing to meet deadlines, and often have to deal with a political environment that could detract from value creation. But we want to create a consulting environment that feels like a creator role, where people are incentivized to build products, pursue new and exciting avenues of research, and most of all, feel like they belong with like-minded people.
Our approach is that if we can make the Collective a place worth working, we will get the best people and make them happy. We believe that this by itself will lend to disproportionate value creation; and by distilling our collective insights and standardizing our offering, we can mitigate the operator problem.
In a nutshell, it’s why we do this.
Quant Collective has helped organizations sustainably generate billions of dollars through advanced analytics applications. Click here to schedule a discussion with us!
We're always on the lookout for exceptionally talented data scientists, machine learning engineers, and business analysts. Click here to drop us a line!