Scoring Apollo.io lead lists

How To
,
1

Hi, everyone! It’s my first time trying out Make after I discovered that a very particular workflow I can accomplish manually could seem to be automated. I still have a few questions after learning the basics of Make so I’ll try to make this as digestable as I can to get some of your insights.

I’m a seasoned tech sales professional and I generate lead lists from Apollo.io to carry out our businesses, but their filters aren’t as precise as I need and a handful of these leads could not be good fits. My lead scoring system uses (impressively well) DeepSeek AI to generate a .csv file output with multiple columns like first and last names, company name and a fit assessment that could be either a bad fit, moderate fit and strong fit; for each lead. Sometimes I only pick strong fits, or moderate fits and strong fits combined. So, to put it all together and get the result I need, I pull all csv files into a Google Sheets’ spreadsheet. Then I filter those by their fit assessment creating what I call a scored list.

My goal isn’t to necessarily upload this to any other software tool for this scenario, but just to accomplish this fundamental part of the workflow. Here’s the basic framework that I visualize for accomplishing this:

  1. Google Drive (Watch All Files) >
  2. Read new unfiltered csv lead list (if it’s a new list) >
  3. Run DeepSeek to score leads and create a personalized line for each lead>
  4. Put together in the same spreadsheet, a sheet of unfiltered leads, a sheet of fit assessments, and create a third sheet with filtered results with an added column for a personalized line.

My biggest questions are related to data processing for the large list of leads that it really is. I often analyze the leads in bulk and get one, big final csv file.

  • Passing a csv file to DeepSeek via make is just as sending this?
First,Last,Company,Website
  • What is the suggested way to process this kind of data? Transform csv’s into readable data for Make? I’ve seen i’m supposed to use arrays but still not familiar with it.
  • I also saw it isn’t possible to modify or work direclty with csv files from Make as I would just normally do using Google Sheets, right? Or I can just work with both my unfiltered list and the final, scored list as also a csv without using Google Sheets at all? Manually, I would have to.

So, this is it. I hope I didn’t miss anything. It seemed nobody ever did this so I decided to create this post. My goal is to just have a final csv file, but I could use Google Sheets later for visualization of the results of course.

Thank you

3

@matheus1 If I understand you well
You when All data to get score by deepseek before it processed to google seet for final processing, but what I think you need is to first connect Appolio to make with there API access it will be more easier to work with than using CSV to googlesheet to make. it will reduce your stress.

Then we can work on how to use arrray to connect filter you scoring needs

4

Thanks for responding, Emmie. Just on a side note: I don’t use Apollo API because we have a cheaper scraper. Apollo could be a bit expensive for us to have leads that aren’t all good fits from the get go, but data is still coming from Apollo’s databases and following their usual format.

I took some time to think of a better way to automate this and realized I was going about it the wrong way. I realized it’s much cheaper and faster to only instruct the AI to give me a column of fit assessment in the particular order of the file submitted for each row and then I’ll import that to a Google Sheet where data manipulation can be a bit easier. I just simply use a FILTER() function in Google Sheets and by adding a new sheet together with an unscored list, I’ll get the result I’m looking for.

The real wrong way to do this is to tell AI to give me the whole list with an added column which would consume 100x more tokens and take much longer to complete.

So far this is how it’s looking to me:

  • Watch Files in Drive >
  • Unless it isn’t a new file, run it through the AI and get a 1 column csv output >
  • Import unfiltered csv list into Google Sheets and add column with Make to the file all at once >
  • Create a new sheet in the spreadsheet to filter results by the new column >
  • Export scored list as csv to a new folder in Drive called “Scored” >
  • Finish

This is what looks doable to me and I’m gonna get dirty on building this.