Intelligence Module · Propensity to Buy

Growth plan +

Stop guessing who will buy. The model already knows.

Propensity modeling scores every customer from 0–100 on their likelihood of purchasing in the next 30 days. Send campaigns to the customers who will actually convert — not your entire list.

Propensity Scores · Live● Updated 4h ago8,432 customers scored

31% Low

24% Below Avg

18% Avg

17% High

10% Very High

CUSTOMER	SCORE	BAND	LAST PURCHASE	KEY SIGNAL	ACTION
Sarah M.	94	Very High	3 days ago	3 purchases in 14 days	Target now →
James T.	87	High	8 days ago	Browsed 6× this week	Target now →
Priya K.	61	High	14 days ago	Email click rate ↑	Include in campaign
Daniel W.	34	Below Avg	31 days ago	Declining frequency	Nurture only
Emma R.	12	Low	67 days ago	No recent engagement	Exclude from paid

Top 10% converts at 11.4× the rate of the bottom 10%

Without scores, your budget spreads equally across all groups

THE PROBLEM

You're sending your best campaigns to your worst prospects.

Without propensity modeling, every campaign goes to everyone. Your $4,700 lifetime customer and your $47 one-time buyer get the same email, the same discount, the same retargeting impression. You optimise for open rates and click-throughs — metrics that feel like progress but don't tell you whether the right people are receiving your message.

They're not. A fraction of your customer base accounts for the majority of your revenue. The rest are noise. Right now, you can't tell them apart until after you've spent your campaign budget trying.

SPRAY AND PRAY

50 customers targeted

Campaign budget: $500

Conversions: ~8 customers (16%)

Wasted spend: ~$330

PRECISION TARGETING

17 customers targeted (score 61+)

Campaign budget: $170

Conversions: ~7 customers (41%)

Budget saved: $330 · Same result

Same revenue outcome. 66% less spend. The only difference: knowing who's ready to buy.

THE COST

Spray-and-pray has a price tag. Here's what it costs.

40–60%

Of campaign budget targets customers who won't convert

In a typical e-commerce store, 40–60% of the customers receiving paid campaigns have a propensity score below 25. They haven't engaged recently, their purchase frequency has dropped, and the model already knows they won't buy. But without propensity scoring, you don't. So you pay for every impression and every send — for nothing.

15–25%

Margin lost on discounts sent to customers who'd buy at full price

Your highest-propensity customers — those scoring 80+ — are already planning to buy. They don't need a discount to convert. But because you send the same offer to everyone, you hand away margin to customers who would have paid full price. Propensity scoring combined with Discount Sensitivity tells you exactly who needs an incentive and who doesn't.

14×

Conversion gap between your best and worst propensity segments

LiftSignals customers see top-decile customers (score 81–100) convert at 11.4% versus 0.8% for the bottom decile — a 14× difference. That gap exists in your store right now. Without propensity scores, you're deploying the same budget across both groups. With them, you concentrate on the 14× group and watch your ROAS climb.

THE SOLUTION

A 0–100 purchase probability score. Updated daily.

LiftSignals Propensity to Buy builds a machine learning model on your store's historical order data — combining purchase recency, frequency, browse behavior, email engagement, AOV trends, and seasonal patterns — to calculate the probability that each customer will purchase in the next 30 days. Every score updates every 24 hours.

Six behavioral signals feed every score

Purchase recency · order frequency · browse session depth · email open + click rate · AOV trend · days since last engagement. The model weights each signal dynamically based on what historically predicts conversion in your specific store — not a generic template applied to everyone.

Scores recalculate every 24 hours

A customer who browses three times in a week moves up. One who stops engaging moves down. Scores always reflect yesterday's behavior — ready for today's campaign decisions. No stale data, no manual refresh required.

Five bands with a default action for each

Scores map to five bands — Very High, High, Average, Below Average, Low — each with a recommended campaign action built in. Export any band to Klaviyo, Meta, or Google in one click. Segment definitions are yours to customise.

SCORE BAND REFERENCE

Every customer scored. Every band has a playbook.

Five bands. Five strategies. No more campaigns built on guesswork.

SCORE	BAND	% OF CUSTOMERS	AVG CONVERSION	CAMPAIGN ACTION	AD STRATEGY
81–100	Very High	~10%	11.4%	Target at full price · no discount needed	Full bid retarget · Lookalike seed
61–80	High	~17%	6.8%	Include all campaigns · small incentive optional	Broad retarget · email sequence
41–60	Average	~18%	3.2%	Nurture with value content — no hard sell	Low-bid retarget only
21–40	Below Avg	~24%	1.1%	Email re-engagement only · exclude from paid	Suppress from all paid
0–20	Low	~31%	0.8%	Full suppression · do not spend	Exclusion list — all channels

Conversion rates are averages across LiftSignals customer base. Your store's distribution varies by category and lifecycle stage.

Propensity to Buy combines with Discount Sensitivity to find the optimal action per customer: High Propensity + Full-Price Buyer = target at full price. High Propensity + Discount Seeker = target with small incentive. The AI Insights brief surfaces these compound recommendations automatically.

HOW IT WORKS

Three steps. Zero manual work.

Your history trains the model

LiftSignals ingests your complete order history, email engagement data, and behavioral signals on first connection. The model identifies which combinations of signals predict purchase in your specific store — weighted by recency, seasonality, and customer lifecycle stage. Stores with 6+ months of history see the highest accuracy.

Runs on first connection

Every customer gets a score

The model scores every customer simultaneously — a 0–100 propensity score plus a band assignment. Scores update every 24 hours as new behavioral signals arrive. A customer who starts browsing heavily moves up. One who goes quiet moves down. The segment you export today reflects yesterday's behavior.

Updated every 24 hours

Export bands directly to your campaigns

One click exports any score band to Klaviyo as a list, to Meta Ads as a custom audience, or to Google Ads as a customer match. Your Very High band becomes your priority email sequence. Your Low band becomes your suppression list. No CSV. No manual upload. Available on Growth plan and above.

One-click export · Growth plan

CAMPAIGN PLAYBOOKS

The right campaign for every score band.

81–100Very High Propensity

~840 customers11.4% avg conversionTop priority

These customers are actively primed to buy. They've purchased recently, their browse frequency is elevated, and their email engagement is strong. This is your most valuable campaign segment this week — the customers who'll convert at the lowest cost per acquisition. Don't dilute the signal with discounts they don't need.

→ Launch a dedicated email sequence within 48 hours — subject line references their last product category or browsed item

→ Retarget on Meta and Google at full bid — do not suppress this segment from any channel

→ Cross-reference with Discount Sensitivity: if they're a Full-Price Buyer, no discount needed. If Discount Seeker, a 10% incentive may accelerate — but don't exceed 15%

→ Export to Klaviyo → "30-Day Propensity VIPs" and trigger your highest-quality sequence

→ Use this segment to seed lookalike audiences on Meta — their profile is your best acquisition signal

Avg conversion on Very High band with targeted sequence: 11.4% vs. 0.8% for an untargeted full-list send

0×

conversion rate gap between top and bottom propensity decile

0.0%

average conversion rate on Very High band with a targeted sequence

reduction in campaign spend when targeting top 2 bands only — same revenue output

24h

score recalculation cycle — always reflects yesterday's customer behavior

GET STARTED

Find out who's ready to buy from you this week.

Connect your store. LiftSignals scores every customer automatically and surfaces your highest-propensity segment ready to target within 24 hours. No data science required.

No credit card required · Setup in 8 minutes · Growth plan and above

Next Module →

Next Purchase Prediction

Predict exactly when each customer will buy again