“What do you do?”
You probably hear that question all the time, and as an individual, you likely have a standard set of answers you give or expect. “I’m an engineer,” “I work in sales,” “I’m a research scientist,” or “I’m a journalist” all give a pretty clear insight into someone’s day-to-day.
But asking a business what it does is far less clear. From niche industries to unique business models to innovative customer experiences, there could be multiple ways of describing efforts as simple as, “We sell goods over the internet” or “We package and ship corn bushels.”
Why is industry classification important?
Since the 1930s, industry experts have refined a framework, now called the North American Industry Classification System (or NAICS), for classifying businesses into standard industry categories.
Before 1938, there was no way to associate census data from different government agencies. For example, production information collected by one agency could not be compared against employment information collected by another.
Over time, industry classification has expanded beyond government and made its way into a range of business processes. Some examples are below.
Insurance: Industry codes influence the risk profile of a business which drives insurance premiums.
Loan Underwriting: Industry codes can influence financing rates. (For example, a sandwich shop generates a different cash flow than a jewelry store.)
Small business classification: The Small Business Administration (SBA) sets standards for what is considered a “small” business based on industry codes. Certain government contracts are designated for small businesses only.
Industry classification has become a major factor in the way that companies access capital, revenue opportunities, tax treatment, and other financial services.
What’s wrong with classification today?
With more than 20,000 industry descriptions, there are frequently inaccuracies, compliance challenges, and operational overhead in answering the straightforward question, “What do you do?”
This problem is even more pointed in regulated industries, where companies manage a list of high-risk or prohibited industries. For example, a payment processor in the U.S. might choose not to engage with merchants in the cannabis, firearms, or adult entertainment industries, among others.
To understand if a business is operating in one of these spaces, you have two options:
Rely on your customers to self-identify their industry.
Companies design a manageable picklist for customers or partners to self-select their industry. The company then uses the set of industries to inform compliance and business rules.
This approach places an incredible amount of trust in a company’s clients to disclose the correct industry. There are clear risks to self-disclosure of customer industry.
Rely on historical industry classification codes, provided by incumbent vendors.
Companies like Dun and Bradstreet and LexisNexis have built large, historical datasets on past industry classifications. However, these companies have fundamental problems around data quality and rely on self-reported inputs as well.
Here are a few real-life examples:
- Coinbase — code 518210 — Data Processing, Hosting, and Related Services
- Pax Labs — code 518210 — Data Processing, Hosting, and Related Services
- PornHub — code 517110 — Wired Telecommunications Carriers
Incumbent vendors also have poor data coverage. After a year in business, there is no record of Middesk in LexisNexis (the industry standard for business risk). In fact, we found coverage gaps as high as 60% when pulling samples of industry codes from incumbent vendors.
Without a reliable solution, operations teams review websites, Yelp, and Google in a best-effort guess to determine the industry of new clients.
How does True Industry work?
True Industry allows companies to manage industry classification at scale.
First, Middesk aggregates information about a business based on the web presence of that company. Examples could include content from their website, social profiles, news articles, and other search results.
Next, we leverage our machine learning model to classify the content that we’ve aggregated about that business. We assign confidence scores to the industries that we believe are most likely related and return a response.
We’re starting out by flagging high-risk or prohibited industries including:
- Adult Content and Services
- Cannabis and Cannabis Related
- Virtual Currency
- Regulated or Illegal Services or Activities
- Gambling or Gambling Activities
- Drug Paraphernalia
- Pseudo Pharmaceuticals
- Other High-Risk Drugs
- Money and Legal Services
- Investment and Credit Services
- Financial Aggregation
- Social Media Activity
- Other High-Risk Businesses
Using our examples above, True Industry provides the following classification:
For businesses that implement True Industry into their customer onboarding programs, our goal is to provide the following:
- Accurate and specific industry categorization — that gives you a clear answer to the question ‘what does this business do?’
- Real-time data that is always up-to-date — For example, a tobacco shop may decide to begin selling cannabis products, a change in which we can notify our clients of in real-time.
- Taking control of risk — don’t leave it to your customers to self-identify industry or historical out-of-date datasets.
While we are starting with the classification of high-risk industries, True Industry will be expanded in the coming months to manage a broad classification of businesses to capture the trends and specific nuances of today’s economy.
How do I start using True Industry?
True Industry is available starting today. To get started, submit a few pieces of information HERE.
We’ll get you set up with dashboard and API access and you’ll be able to start classifying businesses in no time.
As always, don’t hesitate to reach out to email@example.com with feedback or questions.
The Middesk Team