Finding companies that have an interest in buying your product or service is a big part of B2B business. As you can no doubt imagine, this process would be a slow trudge if you had to manually search for companies online (or in a phonebook!) and review their viability one at a time.
Thankfully, the introduction of industry taxonomies such as the Standard Industrial Classification (SIC), which groups companies based on their economic activity, largely removed this potential pain point, as it allowed businesses to find dozens, hundreds, or thousands of companies that all, on paper, shared an industry.
However, as you might have learned from life experience, the removal of one point of friction often leads to the emergence of another. Standard Industrial Classification (SIC) codes are, unfortunately, relatively outdated, static, and oftentimes quite imprecise.
That’s not to say that they’re obsolete. For some companies, SIC codes work just fine, but, based on the conversations we’ve had, it seems that, for a growing number, they’re not quite hitting the mark. This was also the case for us.
So, here at Vainu, we stopped using SIC codes altogether and replaced them with labels based on our industry taxonomy, Vainu Custom Industry (VCI), a taxonomy we specifically created to cover the drawbacks we identified when it came to using SIC codes.
Multiple industry labels
As the title of the blog suggests, we’ll be focussing on a very specific element on VCI in this post—confidence score filtering—so I’d recommend going to our VCI blog post if you want to know more about VCI generally. But, you can do that later because, for now, all you need to know is that VCI assigns companies multiple industry labels, whereas standard classification systems only allow for one.
To illustrate the difference between these two approaches, here’s an example:
- Company: Salesforce
- Standard Industrial Classification: “Software”
- Vainu Custom Industry: “CRM”, “Software”, “Enterprise Software”, “SaaS”, “Information Technology”, “Marketing Automation”, “Cloud Computing”, “Sales Automation”, “Salesforce”, “Customer Service”
As you can see, VCI provides multiple labels, and because of its ability to provide multiple labels, it generally does a better job of communicating to you what type of company Salesforce is and which industry verticals it operates in. You might surmise that SIC codes provide simplicity (one label) whereas VCI labels provide precision (multiple labels).
Confidence scores for industry labels
It’s important to note that all VCI labels are predictions based on our AI model. And, since they are predictions, there’s also a prediction metric that comes along with the actual label. That prediction metric is called confidence score.
Let’s look at some of those Salesforce labels as examples:
- CRM: Confidence score 0.583; confidence threshold 0.24
- Software: Confidence score 0.630; confidence threshold 0.3
- SaaS: Confidence score 0.510; confidence threshold 0.3
- Salesforce: Confidence score 0.337; confidence threshold 0.21
Salesforce is assigned the four labels above because the confidence score for each prediction is higher than the confidence threshold for that label.
Now, a confidence score does not equal probability. In other words, we’re not saying that there’s a 58.3% chance that Salesforce is operating in the CRM space. What we’re actually saying is that we’re very confident that Salesforce is operating in the CRM space, and that can be seen by the confidence score, 0.583, being significantly higher than the confidence threshold, 0.24.
The confidence score of a prediction can be very useful when it comes to sorting and narrowing or potentially expanding the total number of search results. If you wanted to find companies that operate in the field of CRM, you would, for example, simply select “CRM” VCI in our platform, and you’d get back more than 127,000 companies.
These are all of the companies that have been assigned the “CRM” label and have a confidence score over the defined minimum threshold for that specific VCI label. That means that not all of the results will necessarily be equally relevant.
Let’s use seven companies as examples. We already mentioned Salesforce, so that’s one. Then let’s pick two more CRM vendors, like Zoho and Zendesk, two software platforms that integrate nicely with CRMs, such as Calendly and Zapier, an email marketing platform, Mailchimp, and, for illustrative purposes, Accenture, which provides CRM consultation among many other things.
Below you can see the CRM confidence scores for each of these companies. The confidence threshold is the same (0.24) because confidence score thresholds are VCI specific, in this case, specific to the "CRM" VCI label, not company specific.
Salesforce: Confidence score 0.583; confidence threshold 0.24
Zoho: Confidence score 0.539; confidence threshold 0.24
Zendesk: Confidence score 0.515; confidence threshold 0.24
Calendly.com: Confidence score 0.332; confidence threshold 0.24
Zapier: Confidence score 0.326; confidence threshold 0.24
Mailchimp: Confidence score 0.278; confidence threshold 0.24
Accenture: Confidence score 0.2401; confidence threshold 0.24
All of these companies have a CRM confidence score above the defined threshold, which means that, by default, our “CRM” VCI search returns all of them. However, it might mostly be results like Salesforce, Zoho, and Zendesk that a user expects to see.
Seeing Accenture pop up in the search results might come as a surprise, even though it’s a global partner for Salesforce and delivers up to 50,000 hours of Salesforce training annually, and has the highest number of Certified Technical Architects in the market.
Also, Zapier and Calendly might be somewhat surprising results in the list. Zapier is a powerful connector, but it’s not a CRM. Calendly is an appointment scheduling software but, again, not a CRM. However, both platforms are often used together with the CRM, and they both list several CRM vendors as integration partners on their website.
The point is this: If a user were to apply the “CRM” VCI label to their search, expecting to only get back search results for CRM vendors and not software that connects to CRMs or CRM consultants, they’d be in for a surprise.
Confidence score filtering
The good news is that with Vainu’s new search platform, the user can decide how many and how “good” they want the results to be since our new search platform supports searching based on a confidence score.
As mentioned, when searching for the “CRM” VCI, the default threshold returns all those 7 companies mentioned above. However, if the lower threshold was increased from 0.24 to 0.3, Accenture and Mailchimp wouldn’t be part of the results anymore. So, in practice, as the confidence threshold is increased, the number of results goes down, and the relevance of the companies to that specific label goes up. For example, if the threshold was increased to 0.50, only Salesforce, Zoho, and Zendesk would still appear in the search, and they’re all CRM vendors.
This feature was designed based on requests from SalesOps and RevenueOps, as well as SDRs and BDRs, and it makes sense why. They’ll quite often have a very clear Ideal Customer Profile (ICP) and having the ability to narrow down the search and find the most relevant companies to that industry label when needed is crucial for them.
How can I access confidence scores?
Besides being available on our platform, and thus through our CRM Connector, Vainu Custom Industry confidence scores are also available through our API. So, if you wanted to sort those 127,000 CRM-labeled companies and import the data into your CRM, that’s now very easy to do.
Our global offering also integrates natively with HubSpot through our Connector, and, if you wanted to take advantage of that native integration, our recommendation would be to add a new company property called “Vainu Custom Industry” to your HubSpot so that all this industry vertical data is visible and searchable inside your CRM and marketing automation platform.
Hopefully, you now know quite a bit about confidence scores and how you can apply them to find the most relevant companies for you. If you want to hear more about our Vainu Custom Industry or our broad selection of other available data fields, and discuss their applications in more detail, then please feel free to contact us via our chat. You might also consider signing up for our free trial, which would allow you to access our entire company database for a limited time.
