Now live · Tie Predict

Know which shoppers are ready to buy. Before you send.

Tie Predict scores every shopper on your Klaviyo list, and thousands more your ESP cannot see, for purchase intent. Every day. So your best sends reach the right people while intent is still hot.

Identity-backed purchase intent Daily scores in Klaviyo Holdout-tested results
Plugs into the stack you already run
Shopify Klaviyo Postscript Meta Google
The shift

The engagement era is over.

Email built on engagement history (who opened last month, who clicked last week) was designed for a world where your opted-in list represented your total purchase intent. That world has ended. Engagement scoring tells you what a shopper did. It cannot tell you what they are about to do.

Where your purchase intent actually lives
Engaged
Suppression list
Anonymous traffic
The slice you score todayWhoever opened in the last 90 days. A shrinking fraction of your demand.
The list you stopped watchingSuppressed shoppers who failed an engagement filter on one device, while actively browsing on another.
The visitors you never see60 to 80% of site traffic is anonymous. Your ESP has never met them.
  • 01
    Your active segment is shrinking

    Apple Mail privacy, signal loss, and list fatigue erode engagement metrics every quarter. The segment you send to keeps getting smaller, and less representative of who is actually about to buy.

  • 02
    Absence of engagement is not absence of intent

    A shopper browsing on a new device looks "unengaged" to your ESP because it has never seen that device. They are not cold. They are invisible.

  • 03
    The intent gap grows on its own

    The population of high-intent shoppers who do not appear in your engaged segment expands over time. Every program running on engagement scoring alone has one.

54%
of mobile impressions now lack identifiers
Comscore, 2026
$14.19
average Meta ecommerce CPM, up 20% year over year
Sovran.ai
+14%
year-over-year rise in DTC bankruptcy filings
5W Research DTC Graveyard, 2026
In their words

The operators already running it.

Retention and growth leaders at DTC brands on what changed when they swapped engagement guesses for identity-backed intent.

"Instead of blasting broad segments, Tie Predict helps you hit the right people at the right moment, so you're driving better engagement, higher LTV, lower unsubscribes, and improved account health."
Marina Carroll Marina CarrollHead of CRM Strategy, New Standard Co
"We tested Tie Predict head-to-head against another solution and ended up capturing the same revenue while sending to 20% fewer people. Less volume, same return, and a list that's actually healthier for it."
Kyle Turadek Kyle TuradekSr. Director, Growth and Ecommerce, Caraway
"Every brand is running a similar playbook. Predict pulls you out of that and into something actually predictive. We're not just working with purchase history. We're getting scored on real behavioral signals, updated daily. It makes AI actually useful, instead of just marketing talk."
Connor MacDonald Connor MacDonaldCMO, Ridge
"Predict pulls from signals our Klaviyo data just doesn't have, and you feel the change fast. We saw an 18%+ lift in revenue per campaign."
Matt Fey Matt FeyMarketing Director, Portland Leather Goods
Two ways to score a shopper

Stop scoring the past. Start scoring intent.

Engagement scoring · backward-looking

What they did last month

  • Looks only inside your ESP. The list is the ceiling.
  • Mistakes a quiet inbox for a cold shopper.
  • Cannot see the 60 to 80% of traffic that is anonymous.
  • Tells you history, never the next 24 to 72 hours.
Identity-backed intent · forward-looking

What they are about to do

  • Starts outside the list, across every shopper who touched your brand.
  • Reads cross-device browsing your ESP marks as dormant.
  • Matches anonymous visitors to identity, then scores them.
  • 55+ signals per shopper, refreshed every day.
Play · 60 seconds

Ready to buy? You decide.

Here are eight shoppers and everything engagement scoring can see about them. Three will buy this week. Pick the three you would send your best offer to. Then run Tie Predict and see who was actually about to buy.

1 Pick 3 by engagement
2 Run Tie Predict
3 See who was ready
Selected 0/3
How it works

Three intent scores. One you already have: Klaviyo.

Every morning, Predict scores your full identified audience, your opted-in list, re-identified anonymous visitors, and suppressed profiles with cross-device signal, on three dimensions. Each lands on the shopper's Klaviyo profile as a property you can segment on today.

9

Purchase Intent

How likely this shopper is to buy in the near term. Your primary signal for campaign and flow segmentation.

7

Open Intent

How likely they are to open if you send. Protect sender reputation by holding back low-intent contacts.

8

Click Intent

How likely they are to click through. Built for product launches and time-sensitive sends.

1

Tie identifies your visitors

The identity graph recognizes anonymous shoppers on your site and matches them to known profiles. This is the Tie ID foundation.

2

The engine scores daily

Each morning Predict reads the prior 24 hours of behavior across devices and sessions, and scores every profile.

3

Scores sync to Klaviyo

Three properties land on each profile: tie_purchase_intent, tie_open_intent, tie_click_intent.

4

You segment on intent

Swap "opened in last 90 days" for "purchase intent 7 or above." Your existing flows do the rest. No new dashboard.

It is a different input set, not a better guess.

Engagement scoring asks what a contact did inside Klaviyo. Predict asks what a shopper is doing right now, across every device and session, then enriches it against Tie's identity graph.

55+
signals per shopper
280M+
US consumer profiles
300+
attributes per profile
On every Klaviyo profile
// segment condition
tie_purchase_intent >= 7
tie_open_intent    >= 6
tie_click_intent   >= 5
// refreshed daily, 1 to 10
Under the hood

55+ signals. One identity.

Every shopper is resolved to a real person, then scored on dozens of live signals across four families. The engine refreshes all of it, every single day.

Proof

Holdout-tested, with each brand's own data.

Not modeled projections. These are outcomes measured against a control group, on the brand's real list. Every number here traces to a published Tie case study.

Sharper ImageTie Predict
+41%
revenue per recipient
2x
placed-order rate
+36%
click-through rate
Tie has played a big part in bringing our email program into the modern age and giving us the ability to work leaner while feeling far more confident in our sends.
See the proof microsite →
Cozy EarthPredict + ID
15x
return on investment
+40%
open rate
+30%
click rate
Tie helped us unlock revenue we didn't even know existed without hurting our email deliverability.
TUSHYPredict + ID
$110K+
incremental revenue
6x
return on investment
Tie runs quietly in the background and delivers clear ROI. That's what we needed, something scalable, trustworthy, and easy to manage.
Beekman 1802Tie ID
$520,947
net revenue recovered
108,647
shoppers reactivated
We seamlessly reactivated over 108,000 shoppers, generating over $520,000 in revenue, without disrupting our deliverability or existing flows.
+40%open rate, 20% fewer sends

Caraway sent 20% fewer emails and saw 40% higher opens and 30% higher clicks. Predicting intent lets you send leaner and convert harder, which protects sender reputation without giving up revenue. And in one program, thousands of active shoppers were surfaced from a suppression list of over 100,000. They were never gone. They just needed a scoring tool that could see them.

Who it is for

Built for DTC brands with more demand than their list can see.

The fit

DTC ecommerce, $5M+ in annual revenue
100K+ US monthly unique visitors
On Shopify or Shopify Plus
On Klaviyo, or Braze, Bloomreach, Iterable
Fashion, beauty, home, food and beverage, supplements, jewelry
The retention marketer

You own the flows. You feel the list shrinking.

  • Know which subscribers are most likely to buy at any moment
  • Recover high-intent shoppers buried in your suppression list
  • Stop oversending, without shrinking the program
The head of marketing

You answer for channel ROI. And for BFCM.

  • Prove email is driving measurable incremental revenue
  • Send leaner, convert harder, lower blended acquisition cost
  • Install intent scoring before peak season locks
Get started

Live in your stack in one to two weeks.

No replatforming. Predict layers on top of the Shopify and Klaviyo you already run. Entry is a proof of concept, measured with a holdout, so you see the incrementality on your own data before you commit.

From kickoff to scored profiles

  1. Install the tracking layerStandard pixel and API on your Shopify or Shopify Plus storefront.
  2. Connect KlaviyoOne API key. No new platform access beyond your existing Klaviyo admin.
  3. Scores begin syncingDaily intent scores land on profiles within 24 to 48 hours of go-live.
  4. Define the test and the holdoutYour onboarding team helps you set the Predict-scored segment against a control.
  5. Read the incrementalityAt the proof-of-concept close, you get a holdout-tested report on your brand's own results.
24 to 48h
to first scores
1 to 2 wks
to fully live
90 days
proof of concept
The proof offer

Start with proof, not a pitch.

You run a 90-day proof of concept against a holdout. At the end, you have a measured read of the incremental revenue Predict drove, on your list, not an industry benchmark.

  • White-glove onboarding and a named CSM
  • No replatforming, no new dashboard to learn
  • Measured against your own data, with a holdout
Book a demo →
Go deeper

More on identity-backed intent.

The thinking behind Predict
Proof and reports

See your intent gap.

In a 30-minute demo, we will show you how many high-intent shoppers your engagement scoring is missing right now, and what it is worth to reach them before they leave.

DTC brands lock their stack in Q3, before BFCM. The brands that install intent scoring now go into peak season seeing the shoppers everyone else is guessing about.