I was having a conversation with a friend who is an expert in AI yesterday about how I scrub data and if there was a better approach. I wanted to ask this group if anyone has found a way to take data and scrub it to achieve a 20% return rate with better leads?
I remember @Jarenb7 on one podcast stated that he and his team can take days to scrub data, am trying to not be ruled by my business.
Any thoughts?
Hope to be able to meet some of you at the conference this week in PHX.
@gablesland sounds like an interesting idea. I don’t know of any ways to deploy AI for this purpose… but that doesn’t mean it doesn’t exist (or couldn’t be created).
@Jarenb7 would have to jump in to add clarity on what he removes, but I remember him saying he removes properties that show up with wetlands in the wetland mapper, and I think there were some other criteria he uses to remove them.
You don’t necessarily have to follow his logic though. In fact, you might actually increase your response rate if you do mail to these people who have properties with issues… you’d just have to be willing to deal with those issues if you buy them (or make sure your offers are even lower to compensate for those issues).
For example, if you went out of your way to target ONLY properties in an HOA, you’d probably have a fantastic response and acceptance rate, but many of those properties would also have some baggage to deal with.
@retipsterseth I am trying to emulate you and have the business work from the beach! Just wish I could get off the phone, as seems everyone wants to talk and talk and talk…
I have been doing many trials in the mailings and just wonder if people smarter then me can make the data more valuable. I have heard many large well established land investors say that it is all about large mailings, does it have to be…
@retipsterseth , I wanted to add my 2 cents to the AI discussion. AI models need to be trained with lots of data. They help you find relationships in data that are hard to see. If you want a better response rate you would need a large sample of sent mailers data and would need to mark the good responses you got and then AI could scrub your data.
Alternatively you could speed up your data scrubbing if you can enrich your data and then filter it. Enriching can be difficult to automate for no or little money. Usually good data is expensive to get and some data engineering experience definitely helps (data engineers can earn 120k+ per year).
@brianerlich sounds like you’ve done this before. Can you elaborate on what you mean by enriching data? I’m curious to know what that means exactly.
@retipsterseth, Enriching data is adding information. It can be something simple like adding Lattitude and longitude based on an address or it can be complex like adding wetlands information for each lot. There are data vendors that can help you get a phone number for a property owner, a mailing address, or the name behind corporate ownership.
Once you find the data you want, you need a way of logically joining the sets of data. Both sets may have the property ID and county information or it may be a more complex scenario to get the data joined.
Now that you have everything in one data set you can filter it. The easiest way is using Excel. In Excel you will see a data tab in the ribbon on top. There is a funnel icon that is the filter option. Highlight all of the columns that have data and click the filter. This provides you a drop down at the top of each column that allows you to filter by column values.
Its difficult to be more prescriptive than this, without specific details about what you are trying to accomplish and what all of the data sets look like. There is a lot you can do with data and there are a lot of sources for data that is useful. You can also generate your own data, but its most valuable when its maintained in an accurate state. List vendors may be helpful for this. You may also find data vendors on the internet or by asking around your network.