Kwop › Guides › Find hot leads

How to find the hot leads on your website

By Kwop · Updated 17 October 2026

A hot lead on a website is a visitor who went far toward buying (your pricing page, a plan picked, your checkout), left a way to reach them, and hasn't paid. You can find them with a simple score built from a few behavior signals; Kwop (kwop.io) scores every visitor from 0 to 100 and lists the hot ones, with what each was about to spend once Stripe, Whop or Shopify is connected. Below: the intent signals, a spreadsheet score, Kwop's exact formula, and what your checkout leaves behind.

What is a hot lead, and how is it different from a warm or cold lead?

A hot lead has shown strong buying intent and can be reached; a warm lead is interested but not close to paying; a cold lead showed little intent. On a website, intent is what people do: how far they go toward paying, and whether they come back.

In Kwop, these words mean two slightly different things:

• A hot lead on the Follow-ups list has a buying-intent score of 65 or more out of 100, a known email and no payment. Visitor pages label the score Ready to buy (65 or more), Interested (40 to 64) or Just looking (under 40).
• The Hot, Warm and Cold label of the visitor brief (the "AI analysis" panel on a visitor's page, Pro) goes by funnel step only: Hot reached the last step before paying (usually checkout) and left; Warm reached the second step or further, such as pricing; Cold is everyone else. A Hot visitor with no known email isn't on the follow-up list.

Which behavior signals show buying intent?

The strongest signals are the ones closest to paying: reaching checkout, picking a plan, and viewing pricing, especially on a return visit. An email matters for another reason: without it, you can't follow up at all.

Not every signal should add points: rage clicks and failed promo codes change what you write (an apology, a fix), not who you write to first.

A score you can build in a spreadsheet, without a CRM

One row per person you can track, one column per signal, one formula. These weights are an example, not a standard: adjust them once you see who actually pays.

=IF(G2=1, "Paid", 10*B2 + 20*C2 + 35*D2 + 15*E2 + MIN(20, 10*(F2-1)))

With these example weights, 60 or more with an email is hot, 30 to 59 is warm, and under 30 is cold. The maximum is 100: someone who did everything and came back at least twice.

The hard part isn't the math, it's joining visits to people. Analytics gives you counts, not names; Shopify's abandoned checkouts and your email list give you names, not visits. Clarity's Identify API can tag recordings with an email (hashed in the browser), but keeping the sheet current by hand is where this usually breaks down.

How does Kwop score every visitor from 0 to 100?

Kwop's buying-intent score is mostly how far a visitor got in your funnel, plus small bonuses for coming back, leaving an email and viewing pages; anyone who paid scores 100. This is the exact formula in Kwop's code as of 17 October 2026:

If they paid:  score = 100
Otherwise:
  depth = (furthest step reached - 1) / (number of steps - 1)
  score = 10 + 72 × depth
        + 6 per visit after the first   (up to +16)
        + 8 if their email is known
        + 1 per page viewed             (up to +6)
  rounded, and kept between 1 and 99

The funnel is your site's own. Visit is always first and Paid last; in between, Kwop looks for pricing, sign-up, booking and checkout pages, and for links to Stripe, Whop, Gumroad, Lemon Squeezy or PayPal checkouts. Every hour it re-learns which steps visitors really take, and in which order, once the site has 30 visits in 60 days.

A click on a plan button, on a link to Stripe Checkout, a Stripe Payment Link or Whop, or on any link to a /checkout page or with a ?plan= parameter counts as reaching checkout, even though the hosted checkout itself isn't recorded. In the most common funnel, Visit → Pricing → Checkout → Paid, that gives:

Two things follow. Someone who only read your pricing page tops out at 64, however often they come back, so they never become a hot lead. Anyone who reached checkout and left their email scores at least 67, so they are a hot lead until they pay. With more steps, each is worth less: in Visit → Pricing → Sign up → Checkout → Paid, checkout is worth 64.

A worked example from Kwop's demo site: Victor reached checkout (58), came three times (+12), typed his email (+8) and viewed 8 pages (+6, the cap). Total: 84, Ready to buy.

Demo data from Nova Academy, a fictional store.

What isn't in the score: rage clicks, promo code attempts, time on page, traffic source and device. Rage clicks, promo attempts, source and device feed the explanation of why a lead probably stopped and the text of their follow-up; none of them changes the ranking.

How much money is left at your checkout?

Count the people who reached checkout or picked a plan and didn't pay, and value each one at the price of what they picked; when you don't know the plan, use your average order. For example, 40 people who left a $200 checkout this month left $8,000 on the table. That is a ceiling, not money you can recover: some will buy later, and some were never going to.

For scale, Baymard Institute puts the average documented online cart abandonment rate at 70.22% (50 studies, updated 22 September 2025). On Kwop Pro, with Stripe, Whop or Shopify connected, you get it three ways:

• Left on the table (Revenue page). For the period you pick, every visitor who didn't pay and either reached the checkout step or picked a plan, valued at the exact price of the plan they picked (your store's price at that moment), or at your average order when the plan isn't known. The card shows how many people and what share got an exact price.
• Left at checkout, per plan (Products and plans). For each plan, the price of what visitors were buying when they reached checkout with it, for those who bought nothing afterwards: which plan loses the most at the last step.
• Potential spend, per lead (visitor page). The plans that person picked, averaged and weighted toward their deeper and more recent visits. If they never picked one, Kwop uses your average order over the last 90 days, or else your cheapest plan. "Worth in your pipeline" is that amount times their score read as a percentage (a score of 80 on $500 shows $400).

Demo data from Nova Academy, a fictional store.

To win some of it back, see abandoned checkout recovery.

Who should you follow up today?

The highest scorers who left an email and haven't paid, starting with those you haven't written to. In Kwop the list is already made:

1. Overview: the "Ready to buy, never followed up" card counts the hot leads you haven't contacted.
2. Follow-ups: the Follow-ups page lists your hot leads for the period you pick, hottest first, each with the session replay of their visits and an email written from them. Starter opens your hottest 30%; Pro opens all of them. What to write and when: how to follow up website leads.

You can also ask from Claude or ChatGPT (Pro). Connect the Kwop MCP server and ask "who should I follow up today?": the who_to_follow_up tool returns the people with a score of 65 or more who left their email and didn't pay, hottest first, each with the reason they probably stopped and a suggested next step (10 by default, up to 50). The same question works in Kwop AI, the chat inside the dashboard. Your AI gets the facts of each visit, not the replay video, and the tools are read-only.

Demo data from Nova Academy, a fictional store.

What can't a lead score tell you?

A score tells you who to look at first. It can't tell you who will buy, who an anonymous visitor is, or where they work.

• It's a priority, not a prediction. Kwop's weights are fixed, not trained on your past sales, and the "% likely to buy" shown on a visitor's page is that score, not a measured probability. Use it to decide who to write to first.
• B2B account scoring is a different job. RB2B and Warmly identify the company or person behind a visit and score fit (company size, industry, job title) as well as behavior; their lead-scoring guides don't publish their own weights (checked 17 October 2026). Kwop has no firmographic data or company lookup: for B2B account lists, use one of them.
• No name without an email. Kwop names a visitor only when they type their email in one of your own forms or your site calls Kwop.identify. Anonymous visitors still get a score and a replay. More in how to see who visits your website.
• Hosted checkouts are seen from the outside. Kwop sees the click to Stripe Checkout, Whop or Shopify's checkout and matches the payment afterwards by email. A buyer who paid under a different email isn't linked, so they can stay on your hot list.

Sources

• Baymard Institute, Cart abandonment rate statistics: 70.22% average documented rate across 50 studies, updated 22 September 2025. Checked 17 October 2026.
• Microsoft Learn, Clarity FAQ (updated 8 September 2026): free, no traffic limits, no sampling, rage clicks and dead clicks, recordings kept 30 days. Identify API: a custom ID such as an email, hashed in the browser, to filter recordings. Checked 17 October 2026.
• Google Analytics Help, [GA4] Funnel exploration: up to 10 steps, open or closed funnels, drop-off between steps. Checked 17 October 2026.
• Shopify Help Center, Abandoned checkouts: a checkout counts as abandoned when it stays unfinished more than ten minutes after the customer gave their email; listed under Orders. Checked 17 October 2026.
• Stripe Docs, Recover abandoned carts: checkout.session.expired with a recovery URL usable for 30 days, promotional consent required, sessions expire 24 hours after creation by default. Checked 17 October 2026.
• RB2B, How to identify high-intent leads on your website (7 May 2026): behavioral signals combined with firmographic filters; no scoring weights published. Checked 17 October 2026.
• Warmly, lead scoring guide (1 October 2025): person-level website signals, third-party intent and firmographic enrichment, scored against your ideal customer profile; its point table is a generic illustration, not Warmly's model. Checked 17 October 2026.