Developing answers without data is inherently difficult.
The benefits of working at Thinknum gives me access to millions of private and public companies globally. As I sit back and sift through Thinknum Alternative Data, I get quick answers to my questions. It hasn’t always been this easy.
Our story starts in 2013 with Gregory Ugwi. Operating as a non-agency mortgage strategist at Goldman Sachs, he was among a group of elite analysts. An important aspect of Greg’s job was finding the best available data so that Goldman Sachs could stay ahead. He found that the internet was a great source of information, but gathering and consolidating it was quite the challenge and required ingenuity. It involved scraping data directly from web pages. However, many of his counterparts were afraid of web scraping.
Greg Ugwi — Nobody gave us the data, so we had to get it ourselves and clean it.
After Greg left Goldman Sachs, he joined up with his partner, Justin Zhen. Web scraping took on a new meaning for them. They sought to capture the world’s activity as companies and customers fled online. This activity represented digital footprints for them, which they left behind during their various interactions. This company became Thinknum Alternative Data.
Greg Ugwi — We [Thinknum] crawl the web for data that shows economic activity, and organize it into data models so it can be queried. […] We pull in other entities who structure the web’s data and make it accessible to everyone else, in its raw, unedited form.
This ambitious vision is nothing new. Many of the world’s top companies today share this principle. The way Google, Facebook, and CNN aggregates digital economic activity are prime examples.
Building a product or a great company?
Thinknum got its start by giving demos to close friends.
Greg Ugwi — “For the first 12 months, we believed that people would analyze companies and their data on the platform. 24 months later, we realized people didn’t need analytic features, so we threw them out.”
Instead, Thinknum Alternative Data focused on developing a smooth user interface and API. For the interface, they developed a search engine to make their small but growing database more queryable.
Ironically enough, some of Thinknum’s first useful data services covered brick and mortar industries like real estate and store location datasets.
When specific stores go bankrupt, you can query the competitors in a 1-mile radius, to project who may get the remnant business. On the internet, business activity and major trends converge with these companies, accelerating their online activity of e-commerce, social media, and hiring.
Today, analysts who need to be informed use Thinknum to search for everything that is happening in the world and at a specific company. Thinknum’s job listings dataset shows whether companies are hiring or firing. The Products by Vendor dataset shows aggregated product price fluctuations across brand and category, and employee sentiment data shows workforce outlook on management, company success, and satisfaction.
This alternative data was first consumed by long/short hedge funds seeking fast, timely information, and put it to use before others in their industry. Banks use alternative data to enhance their sell side research, asset managers use the data to validate their investment thesis, while corporations and large tech companies use data to remain market leaders. Data has thereby fueled the business cycle.
If a food delivery company wants to know the location of their competitors’ restaurants, they drill down to the new ones they added last quarter.
Take a major telecom and semiconductor manufacturer who wants to track the fastest growing software companies so they can upsell them new IT cloud solutions. Here, data empowers the sales mechanism for the company, increasing revenue and profits, ultimately impacting the company’s value and stock.
You see, data is essentially an arms race. If you just listen to official company reports, like quarterly filings and stock prices, you’re getting what everyone else is. You need to go a step further to differentiate. It’s what we call Alpha, which is where Thinknum Alternative Data comes in.
Remember WeWork? That shared office space
A story that propelled Thinknum to fame goes back to 2015, when I met Gregory Ugwi and Justin Zhen for the first time.
Justin and Greg were the first to uncover cracks in the “billion dollar empire,” WeWork. By crawling the public directory of WeWork, Thinknum knew when customers would join and leave — predicting customer churn.
They published the insights online for investors, community members, and others to consider. There, they discussed WeWork’s high churn and their high cost of customer acquisition. In 2015, this was in stark contrast to what WeWork’s management described, as they raised $100 billion dollars to an eventual IPO.
Adam Neuman gave them 45 minutes to get out.
Five years later, those same insights underpin the entire collapse of an organization that raised billions on false promises and poor execution.
“Billion Dollar Loser, ”written by Revves Wideman, covers this story at length. Furthermore, a Hulu documentary special is in the works that better articulates their story. Both will speak to Thinknum’s data and how it was instrumental to sounding the alarms and clearing up the smokescreen surrounding Adam Neuman and WeWork.
Back to the Basics
It’s hard to get people to start using new, disruptive solutions, but once they do, they rarely go back to the old way. Alternative data is highly dimensional in that the same “data” can be used to show different insights depending on its application.
Employees: Job listings of companies
Employees: Employees’ sentiment and conversations about work
This data itself is not made for investing or any specific application for that matter. There’s also another benefit in that alternative data is objective. Because the data isn’t meant for you, if you do get a hold of it, you can dodge the spin commentary to get clearer insights.
It’s similar to being a silent, invisible observer during a psychology experiment which removes any bias to the study results due to the observer effect.
To grow, companies must hire and retain strong talent to secure value in the marketplace. When demand is good, you’re gonna ramp up, invest more, and hire more. This is where Thinknum Job listings comes in. What your customers say about you is very important. With Thinknum User Reviews, you can see what customers aresaying about your competitors’ products.
Beyond Meat is a strong brand. Can alternative data fuel McDonald’s ability to come out with its own alternative product? What about supply chains? McDonald’s can use Store Location Data to find out whether Beyond Meat is getting good distribution in the market or find its weaknesses to capitalize on.
The business model of Life Insurance Companies is to get more top agents with books of business to join their ranks. These agents are commonly listed online, making them trackable.
In aggregating online data and making it easy for people to access, Greg recognized an important opportunity:
Greg Ugwi — “Index this canonical data into specific data models that everybody shares. Taking a step further, we build an index that is good enough for our pickiest clients, and easy enough for the most novice clients. If we do this well, we’ll add value in the world.”
In this way, data is pushed, rather than pulled. Alternative data becomes a tool for companies to understand the big players and market levers. The goal is not to recreate the entire web. Companies like Snowflake tackle this problem. Of greater importance is the ability to standardize the naming and access to web data.
Structure the internet to build relationships
This has led to an evolution of products for Thinknum with the launch of KgBase, a unique tool that advances semantic modeling technologies at the enterprise level.
KgBase acts as a visual front end which developers and analysts use to build on top of their company’s data assets. By visualizing creating the data model, they leverage an integrated environment to build workflows, dashboards, and other data applications.
Allow developers to use KgBase across multiple clouds, including Neo4j, Amazon Neptune and Azure. Now, data and business analysts get insights more easily leveraging their familiar tools.
This simplifies self-service business intelligence tools while extending the scalability for any amount of data or number of users.
Users who like graph databases use the power of machine learning without having to learn anything else or without having to move their data anywhere else. With KgBase penetrating at the enterprise level, enterprise developers and analysts need to see what lies ahead now more than ever.