8 of 10 - In Which We Talk About Link Analysis
The Name Behind the Name
In wrapping up my Missing Persons trilogy, I went over a series of records one could use to find people associated with a company. I casually threw out at the end of that post:
I’ll end with these two other ideas when you remain stumped. First, some companies, even private ones, have to disclose information to government bodies; there are businesses that are regulated liked insurance companies and broker/dealers, and there are companies that are government contractors or such and have to provide disclosures. Second, there is “link analysis”, that’s taking what you know about one company and trying to find other companies, and then drawing inferences between the companies.
Man, doesn’t link analysis make you sound smart. What do you do. Link analysis. It’s like you’re one of those computer geeks staring at three screens that the superhero comes to for help. On the TV shows, the links just pop up. Click click click click and our link analysis chart germinates before our eyes. Case solved.
It never goes like that for me. I hardly ever even make a nice link analysis chart with arrows and boxes. In most cases, it’s a lot less about diagraming connections and a lot more important to have connections. The right connections as I caution below. Believe me, you can explain that connection in a sentence or two without making pretty pictures. Like try this set. We believe Joe Fraudster is connected to Scamco because Scamco is at 100 Main Street, Suite 100, and several sources show Mr. Fraudster using this address in other contexts. For instance, the car Mr. Fraudster drives is registered to this address. That’s link analysis, showing how a person is connected to a company even though his or her name does not otherwise appear in company documents.
With link analysis, you are striving to do two things. Find something that will lead to another thing, and find enough things that will show a relationship between two or more things. Your goal is to find that thing.
Follow this process:
Start with subject - company or individual
Identify something about that company or individual
Use that something as your “search terms” to see what else you can find
Put two and two together
Rinse and repeat - identify something else, search again, put new things together
Link analysis requires searching and it requires the accumulation of data. Search a company name through a Secretary of State database. It will contain two things, a name and an address. Now, you have at least two somethings for more searches. Your searches can identify other companies connected to that name, and your searches can identify other companies at the same address. These are not the only things you may find on a company. They usually have a phone number, do other companies share that phone number? Website, it may be possible to find other companies with similar registration information or other companies using the same host. One way is part of link analysis is thinking of all the things that could be out there. The other part is to collect all the things that are out there.
That’s the value of the Ripple Theory. Companies and individuals throw off tons of ripples. Like I mentioned above, drive a car, you need to register it, register it and you provide an address that will likely end up in a public record. When we are doing link analysis, we suck up all the data we can. We will schedule out real estate transactions or litigation filings so we have each name that comes up, each address that comes up, each nugget that we think can either lead to other nuggets or each nugget that can adequately connect A to B to C.
And…and there’s a huge, big, important, vital caveat when doing link analysis. Be aware of the false link. Put it better, be skeptical of any link. There is the danger of guilt by association. You know, pick your political bugaboo. You can do it with Clinton, George W. Bush, anyone, any side, the chart linking them to mafioso, the JFK assassination, Central American Generalissimos. And depending on your inclinations, some of those have to be true. Right? We’ve skipped nodes and networks and the idea that things can be connected with less connection than sharing a phone number. Yet, the more we create networks the more likely our connections will be nebulous. Be especially wary of items that lead to nowhere.
Take CT Corporation. When you do your first search, you identify this as the registered agent. Oooh boy, the links you find. Until you look at CT’s website: Safeguard Your Business With America's Most Reliable Registered Agent Services Trusted by lawyers and more than 300,000 businesses since 1892. In other words, it is doubtful that companies sharing CT as their registered agent are otherwise related. Another thing to understand is the date or history of the potential link. How long ago was a company at an address? Do they still use that telephone number. Using a connection that is not meaningful or using a connection that is stale are two easy ways to negate the value of your link analysis. There are a lot more ways that things that may be connected are in fact not.
It’s eight posts in eight days. Link analysis and its value in finding people connected to companies was tossed out the other day. So I figured I at least had to give the basics by explaining the process of link analysis and the likelihood that it can go awfully wrong. Should be enough for the challenge.