How to find buyers evaluating software

Last Tuesday I watched a founder get a paying user from a Reddit thread that was 47 minutes old. The person had posted a messy, frustrated paragraph about the tool they were using. The founder replied with a specific workaround, not a pitch. That is the entire game.

Finding live evaluation conversations is not a growth hack. It is a timing problem. You are looking for a narrow window where a human being has publicly admitted a problem, named a category, and started comparing options. Most founders are either too late or too broad. They wait for people to arrive at the website, or they set up a social listening dashboard that surfaces every mention of a keyword, including press releases and job postings. Neither gets you into the conversation while it is still warm.

What a live evaluation thread actually looks like

Most public software evaluation is not hidden. It is just buried under noise. The signal you want has a specific shape. I look for three things in a thread before I consider it a live evaluation.

First, the person names a current tool or a specific pain point. Not "looking for note app recommendations" but "I am moving off Notion because it is slow on my phone and I never open it." The specificity is the signal. Vague questions are usually students doing research or people killing time.

Second, there is a time constraint or a trigger event. "My subscription renews in 30 days." "I just started a new project." "I am onboarding two people next week." A trigger event means the evaluation has a deadline. Deadlines are what turn browsing into a decision.

Third, the person is comparing at least two named alternatives, which is the clearest buying intent signal a public thread can carry. "I am down to Linear and Height." "Has anyone switched from Substack to Beehiiv recently?" This is the moment where a well-timed, specific answer can shape the outcome. The person is not just browsing. They are deciding.

A thread with all three signals is rare. Maybe one in forty relevant mentions. But that one thread is worth more than a hundred visitors who bounce. The person is in motion. They have a problem, a timeline, and a shortlist. Your job is to be useful before the shortlist hardens.

Why keyword alerts keep missing the moment

Most founders start with keyword alerts. They set up a tool to watch for "best [category] app" or "[competitor] alternative." Then they get a firehose of noise. Blog posts. Comparison pages. Someone asking a question three years ago. The alerts fire constantly and almost never surface a person about to choose. I wrote about this in more detail in why your keyword alerts keep failing you. The core problem is that keywords describe topics, not intent. A person writing "best CRM" might be a student, a journalist, or a bot. A person writing "I am leaving HubSpot because the reporting is slow" is about to pick something. Keyword tools cannot tell the difference.

Social listening platforms are the second option. They are powerful but built for brand monitoring at big companies, not for a founder who needs users this month. You can build complex queries, but you still end up with a stream of mentions that requires a human to triage. You do not have a dedicated analyst. You have a product to ship. The setup cost is real. The ongoing triage cost is real. And the platform still does not tell you which mentions are worth a reply.

The gap is judgment. You need a system that can read a thread and say: this person is evaluating, this person is complaining, this person is just sharing a blog post. That judgment is what turns a firehose into a short list you can clear before lunch.

A practical method for finding and prioritizing live evaluations

Here is the method I use. It works whether you are using a tool or doing it manually with a few saved searches.

Start with the language of evaluation, not the language of your product. People rarely say "I am evaluating software." They say things like:

• "I am thinking about switching from"
• "has anyone moved from X to Y"
• "my plan renews soon and I am looking at"
• "what do you all use for"
• "X is getting too expensive, alternatives?"

These phrases are the raw material. They appear on Reddit, X, LinkedIn, Hacker News, and in niche Discord and Slack communities. The best threads are often on Reddit, where anonymity lets people be honest about the tool they are leaving. I broke down the structure of these threads in the anatomy of a "what do you all use for..." thread. The pattern is consistent: a specific problem, a named incumbent, and a request for alternatives.

Once you find a candidate thread, score it. Any usable version of an intent score does the same job. I use a simple three-point system. One point for a named pain point. One point for a trigger event or deadline. One point for a named competitor or shortlist. A thread with three points is hot. Two points is worth a reply if you can be genuinely useful. One point is usually noise.

This scoring matters because it forces you to prioritize. You cannot reply to every mention. You should not. The goal is to find the ten threads a week where your answer can actually change the outcome. Ten good replies a week is a marketing channel. A hundred mediocre ones is a reputation problem.

The reply itself is where most founders fail. The instinct is to pitch. "We solve that exact problem, here is a link." That is how you become the spam guy. The better move is to answer the question first. Share a specific insight about the tradeoff between the two tools they named. Mention a limitation of your own product if it is relevant. Then, if it fits, say what you built in one sentence. I wrote about this in how to find customers on Reddit without becoming the spam guy. The rule is simple: be useful first, mention yourself second. The thread is public. Other people with the same problem will read it. Your reply is a permanent artifact.

How this compares to waiting for sign-ups

The usual way a founder learns about a potential user is after the fact. Someone visits the pricing page. Someone signs up and never comes back. Someone opens an onboarding email. These are weak signals. A pricing page visit might be a competitor doing research. A sign-up might be a student. Analytics tell you that someone was aware of you. They do not tell you that a specific person is in pain right now, or what they said about it.

Public evaluation threads are the opposite. They are strong signals from a specific person, in their own words, before they have ever heard of you. You do not know their budget or how many people they are buying for. That is the tradeoff. Analytics give you behavior after the visit. Public threads give you the reason before it.

The way to use both is to treat public threads as the front of the funnel. When you find a live evaluation, you have a name, a problem, and a timeline. That is enough to write one useful reply. If they click through, your product and onboarding take it from there. The thread is the trigger. Everything you already built provides the rest.

The measurement is different too. You are not measuring page views. You are measuring replies posted, conversations started, and people who showed up saying "saw your comment on Reddit." A single thread can be read by hundreds of people with the same problem over the next year. The reply you write today is still working for you in six months. That is the compounding effect of public conversation. It is also why dark social is eating your attribution, and that's fine. The person who found you through a thread will not show up in your attribution reports. They will show up as a user.

The founders who win here are the ones who stop treating social as a branding exercise and start treating it as the place their next users already are. They build a short list of live evaluations. They score each thread. They reply with specific, useful answers. They notice which replies turned into sign-ups. And they do it every day, not as a one-off campaign.

This is the work I do every day. Oetzi exists because I got tired of manually refreshing Reddit searches and X queries to find these threads. It watches for the language of evaluation, scores each thread for how close the person is to choosing, and drafts the reply for a single inbox. It also tracks whether ChatGPT, Perplexity, and Gemini recommend you when people ask, which is a whole other place to be found. But the method I described works with or without the tool. The hard part is not finding the threads. The hard part is having the judgment to know which ones matter, and the discipline to reply like a person instead of a marketer.