What is good research?

I spend 10 hours a week researching for my work at Neatprompts. This is an overview of how I approach my research and why it is important.

After the topic is decided, I need as much raw material as possible around that topic. I ask Perplexity to throw me all the resources it can find. Blogs, Newsletters, YouTube, research papers, etc.

For Neatprompts, I either have to research about top executives or companies. So I usually start with podcasts or any public appearances they have had. I skim through all the resources to understand the volume.

Okay, this depends on the importance of the content I come up with. If the content is generic, I read summaries using NotebookLM.

But for super important newsletters, I read each line and listen to every word. I don’t use AI to summarize and then pick pointers from it.

I do this for two things:

Once I cover the popular podcasts, YouTube videos, and newsletters, I get nerdy and find niche areas where most people won’t get into.

This includes research papers, subReddits, communities, blogs, etc.

Hack: Once you find a niche platform, paste the link into a Google Sheet. Saves time and effort the next time you need it.

If I have to write about entrepreneurship or Indiehacking, I prefer reading stories from Hackernews over Twitter threads. If this sentence made sense to you, you understood what I meant by niche platforms.

Some topics are evergreen, but the operations, methods, and perspectives keep changing. I work in the AI niche, so even one year feels ancient. But if I have to write marketing stories, I prefer older brands to study storytelling and how they measure impact.

So you choose how important relevance and recency are based on what you’re creating.

If I’m talking about a company, I’d love to be its user before anything else. Users’ perspective triumphs over all, so it is important for me to know how I feel as a user before I think as a marketer.

This approach helped me write sponsored posts for the brands that spend thousands of dollars on a single newsletter, and deliver better results than promised.

Always read what the company, its leadership, and its reports say. This includes earning calls, investor reports, PR, etc. If you have a connect, even better. Reach out and ask questions.

Your value increases 10x as soon as you mention you’ve collected first-party data.

In my first round, I mention the sources and write down all the points I think are important. I don’t write keywords, but I actually write sentences by hand. It slows my thinking and generates new angles as I write them.

Then I look at all the sentences and highlight the ones I think will make it to the content. I call these elements.

If my research has 10 points, 4-5 of them would be elements. Then I pick elements, build a narrative and create an outline to weave elements together.

Because I spend way too many hours on research, everything is on my mind already, and I finish writing in less than two hours. Clarity is directly proportional to speed.

All this process is effective with AI in most cases. I follow this regressive manual method only for important, BoFu, paid content. The human touch always stood out.