Most lead-scoring guides are written for B2B software teams chasing demo requests. If you run a store with real orders and real ad spend, the mechanics are the same but the stakes are sharper: every point of your score should answer one question — is this person worth more of my limited retargeting budget and email attention?
This guide gives you the exact criteria, point values worth starting from, and a worked example on an operating store's numbers. It is part of our broader store automation playbooks series.
What "lead scoring" means for an operating store
In an ecommerce context, a "lead" is anyone in your orbit who hasn't bought yet this cycle — a new email subscriber, an abandoned-cart visitor, a lapsed customer. Lead scoring assigns each of them a number based on what they do and who they are, so your sales and marketing automation can treat the 8-point drifter and the 60-point near-buyer differently.
The payoff is budget discipline. You have a fixed ad spend and a fixed number of email sends your deliverability can support. Scoring decides where both go.
The generic guides stop at "demographic + behavioral criteria." That's true but useless without numbers. Below is the version with numbers.
The three categories of lead scoring criteria
1. Behavioral criteria (what they do)
Behavior is the strongest intent signal because it's recent and voluntary. Weight actions by how close they sit to a purchase:
- Opened your last three campaigns: +5
- Clicked a product link in an email: +10
- Viewed a product page two or more times in seven days: +15
- Added to cart, no purchase: +25
- Started checkout, no purchase: +35
A single email open is near-worthless on its own — it can fire from an image preview. Cart and checkout actions are where the real intent lives, so they carry the most weight.
2. Attribute and value criteria (who they are)
Attributes are slower-moving but tell you lifetime worth. For a store, the most useful attributes come straight from order history:
- Past purchaser: +20
- Two or more past orders: +30
- Past average order value above your store AOV: +10
A repeat buyer with a high AOV is worth chasing even when their recent behavior is quiet, because their expected profit per re-engagement is high. This is the signal the B2B templates never translate for stores.
3. Negative criteria (who to stop spending on)
Negative scoring is where most models are thin. Without it, scores only climb and your "hot" list fills with stale contacts. Subtract points for disengagement and low-quality signals:
- No email open in sixty days: −15
- Hard-bounced or marked spam: −40 (effectively removes them)
- Only ever bought once, on a deep discount code: −5
Pair this with point decay — strip the points from a behavioral signal after it ages out (say, zero out that "+35 started checkout" after fourteen days). A score that never decays is a score that lies to your automation.
Worked example: scoring on a real store's numbers
Say you run a store doing 340 orders a month at a $31 AOV, spending $2,800 a month on Meta. Before you can decide a lead is "worth" retargeting, you need to know what an order is actually worth.
Walk the per-order math:
- AOV: $31.00
- Product + shipping cost (Printify/Printful): −$13.50
- Payment processing (roughly 2.9% + $0.30): −$1.20
- Ad cost per order ($2,800 ÷ 340 orders): −$8.24
- True profit per order: $8.06
That $8.06 is the number that makes scoring non-optional. At roughly eight dollars of profit per order, retargeting a pool full of 8-point drifters at the same cost-per-click as your near-buyers burns the margin fast. Computing this true per-order profit — product cost, fees, and ad spend netted out — is exactly the kind of work an AI employee like Victor does across your connected store, supplier, and ad accounts, so the profit figure your scoring model leans on is live, not a stale spreadsheet.
Now set the threshold. Say a lead crossing 50 points gets promoted: added to your paid retargeting audience and dropped into a dedicated high-intent email flow. Everyone below 50 stays in low-cost nurture.
Do the concentration math. If 1,200 contacts are "active" but only 220 clear 50 points, pointing your $2,800 at those 220 instead of all 1,200 raises your effective spend-per-qualified-lead from about $2.33 to roughly $12.73 — you're no longer averaging your budget across people who were never going to buy. The threshold is the whole game; set it where your spend decision actually changes.
Wiring the criteria into marketing automation
Scoring criteria do nothing until an automation platform acts on them. The pattern is the same across tools:
- Collect the signals. Your email/SMS platform tracks engagement; your store tracks orders and browsing. Klaviyo, for example, can build segments from a plain-language description of the behavior you want to target (Klaviyo Help — define segments with AI).
- Assign points with the criteria above, in your automation platform's scoring or property fields.
- Set the threshold that maps to a real action — promote to a paid audience, trigger a flow, or flag for a win-back offer.
- Let points decay so the score reflects now, not six months ago.
Timing compounds the score's value. Klaviyo claims its Personalized Send Time feature drove a "35% lift in click rate" for top campaigns — a vendor-reported number, not a guarantee (Klaviyo — AI for autonomous marketing and customer service). A high score tells you who; send timing helps with when.
On the paid side, your score defines the audience and the platform's own automation handles delivery. Meta says advertisers see a "20% lower cost per result on average" with Advantage+ sales campaigns — again, a vendor average, not a promise (Meta for Business — Advantage+ sales campaigns). Feeding that automation a score-qualified audience instead of a broad one is where your own leverage sits.
If you're still choosing a platform to run this on, our broader inbound marketing automation playbook covers the scoring-capable options.
Where lead scoring goes wrong
- Scoring vanity actions. Opens and impressions inflate scores without predicting purchases. Weight toward cart, checkout, and repeat-view behavior.
- No negative scoring or decay. Scores that only rise fill your hot list with ghosts.
- A threshold that changes nothing. If crossing 50 points doesn't move a dollar or trigger a flow, the number is decoration.
- Ignoring profit. A lead likely to buy a deeply discounted, low-margin item isn't the same as one likely to buy at full AOV. Score toward profit, not just conversion probability.
Can an AI employee run lead scoring for you?
Partly — and the honest line matters here. Victor, the AI employee from PodVector AI, connects to Shopify, Meta Ads, Google Ads, Printify, Printful, Gelato, and Klaviyo, computes your true per-order profit, and delivers reports to your own Google Drive. Every write action it takes is approval-gated: Victor proposes, you approve, then it executes. That approval gate is the industry-standard control, and it's the right one for anything that moves spend.
What that means for scoring: Victor can pull the live order and engagement data your model depends on, surface which segments are actually profitable, and — with your approval — take Klaviyo flow actions on them. It is not a dashboard you stare at; it's closer to an employee you hand the recurring work to. For a fuller picture of what that category can and can't do, see our guide to the best AI agents for business automation.
Want to see your store's true per-order profit and segment economics in one place? Start with PodVector AI.
FAQs
What is the difference between lead scoring and lead grading?
Scoring measures intent and engagement through behavior that changes often — clicks, cart activity, recency. Grading measures fit through slower attributes — for a store, that's order history, AOV tier, and repeat-buyer status. Most working models blend both into one number, which is what the criteria categories above do.
What point threshold should I use to call a lead "hot"?
There's no universal number — the right threshold is the score at which your spend or flow decision actually changes. Start by listing what happens when a lead qualifies (enters a paid audience, triggers a flow), then set the threshold so roughly the top ten to twenty percent of your active list clears it. Tune from there based on how those promoted leads convert.
How do I score leads when I sell low-AOV print-on-demand products?
Lean harder on profit and repeat behavior than on a single purchase. At a thin per-order profit like the one in the worked example above, one discounted sale barely covers acquisition, so weight repeat-order count and above-average AOV heavily, and apply negative points to discount-only, single-purchase contacts.
Do I need a separate tool for lead scoring?
Usually not. Most marketing automation and email platforms have built-in scoring or custom property fields you can run these criteria in. The harder part isn't the tool — it's keeping the inputs accurate, especially the profit side, which means your order, supplier, and ad data all have to agree. That reconciliation is where a cross-tool AI employee earns its keep.
How often should I update my lead scoring criteria?
Review the criteria quarterly and the point decay rules continuously. Behavioral weights drift as your product mix and funnel change, and seasonal buyers distort attribute scores. A model you set once and never revisit slowly fills your "hot" list with contacts who matched last year's store, not this one.