500 free credits, no card. Try Smalt AI →

Sales

How do I find new leads by looking at who my competitors sell to

Build a lead list from competitor signals: buyer-intent data, technology lookups and public customer pages, then filter by fit and check the rules.

Smalt AI team ·

Find the companies that use, evaluate or have left a competitor, then filter them against your ICP (ideal customer profile), the description of who is most likely to buy from you. Buyer-intent signals, technology lookups and the competitor’s public customer pages supply the names. The result is a warm list that still needs qualifying.

How to do it

  1. Collect buyer-intent signals showing who is researching competitors.
  2. Look up which companies run the competitor’s technology.
  3. Read the competitor’s public customer pages and reviews.
  4. Filter every name against your ICP.
  5. Qualify what remains and find the right contact.

Start with buyer-intent signals

Review sites sell data on who is looking at competitors. G2’s Buyer Intent documentation lists signal types including Alternatives, Compare and Competitive. Alternatives means a buyer viewed an alternatives page for you or a competitor. Competitive means a buyer viewed the pricing, profile or compare pages of a competing product in a shared category. Which signals you receive depends on your G2 plan, so check what yours includes first.

G2 also published a guide to using competitors’ buying signals (October 2025) describing three plays: alerts to sales, ad campaigns, and churn defense or displacement. It says these signals typically occur 20 to 30 days before conversion.

Read the numbers in that guide as a vendor’s figures: G2 and its partner Factors.ai looked at 50 mutual customers and counted a website inquiry form submission within 60 days. They report these average conversion rates:

Page viewedConversion within 60 days
Your own product page5.5%
Competitor feature pages1.03%
Alternatives page0.73%
Competitor pricing0.54%
Competitor reviews0.53%

The guide gives no sample details beyond the 50 customers. The figures support a ranking. A visitor to your own page converts about five times as often as one reading a competitor’s feature pages. It is more than ten times as often as one reading competitor reviews. Competitor-page signals give you a warm list, not hot leads, so plan for a low hit rate and a longer follow-up.

Find who runs the competitor’s product

If your competitor is software, you can often see who uses it. Apollo’s guide to finding companies that run a specific technology (vendor marketing, August 2026) makes these points:

  • Customer-facing technology is detectable with tools such as BuiltWith or Wappalyzer.
  • Backend technology usually needs indirect evidence, such as job postings that name the product or published case studies.
  • Layer firmographic filters (company size, industry, location) on top of the technology match.
  • Confirm each account with at least two independent sources.
  • Put accounts that recently adopted or dropped a competitor first.

That last point matters most. A company that just switched away has an open decision, while one that signed a three-year contract last month probably does not.

Read the competitor’s public pages

Competitor case studies, customer logo pages and review text name companies and often describe their situation. Reading them by hand is common practice, not something a source here documents, so treat it as a starting point. Note each company, what they say they use the product for, and what they complain about in reviews. Those complaints become your opening line.

Filter by your ICP, then qualify

Every list above names competitor customers, not yours. Cut it down with the criteria you use for any prospect (company size, industry, location and the problem you solve), then rank by how recent the signal is.

Qualifying means confirming each account with a second source and checking the signal is recent before anyone writes to it.

Skip scraping competitor followers on LinkedIn

Some vendors, such as AiSDR, describe finding people who engaged with competitor LinkedIn pages and filtering them by ICP. Do not collect those people with a scraper. LinkedIn’s help page on prohibited software and extensions says it bans scrapers, bots and automation extensions under Section 8.2 of its User Agreement. It adds that accounts using them risk restriction. Looking through a competitor’s page yourself and noting relevant companies is a different, manual activity.

Doing it with an AI coworker

You can hand the research half to Smalt. Ask it to search the web for a named competitor’s customers, case studies and recent news. Then have it look up each company and add firmographic details for your ICP filters. Ask it to find the right contact at each account, as in this walkthrough on finding a head of operations, add those that pass to HubSpot, and keep the source of each account in a sheet.

Drafted emails wait for your approval before anything is sent, but contacts added to HubSpot do not, so name the exact list. Deciding which accounts deserve outreach stays with you.

What to check before you rely on it

  • Signals are not intent to buy. Even in G2’s own figures, most visitors to competitor pages do not inquire. Someone reading competitor reviews may be a student, an analyst or an existing customer.
  • Technology detection misses things. Backend tools often do not show up at all, and detected tools may be abandoned. This is why the two-source check matters.
  • Customer pages are marketing. A logo on a competitor’s site may be a former customer or one small team inside a large company.
  • The vendor figures are vendor figures. G2 sells intent data and Apollo sells prospecting data. Run it on 50 accounts and count the replies before you commit budget.
  • Check the rules where you sell. Whether you may contact a given person, and on what basis, depends on jurisdiction. Confirm your local privacy and anti-spam rules before sending. Take particular care with personal email addresses and recipients in other countries.

Smalt is an AI coworker that takes jobs like this end to end and hands back the finished file.