Predict Customer Behavior

Easy Machine Learning for Your Website

Identify customers who are likely to convert

Calculate propensity to purchase or churn, product recommendations, best time to contact, and other metrics for your users.

Connect with them

Run a promotion, retarget with ads, make a customized offer, send a campaign. Use with Google Ads, Facebook Ads, Bing Ads, or any other advertising network.

No credit card required

Actionable metrics driven by ML

Get insights into user behavior to enable ML-driven scoring, targeting and personalization. Run pre-built or custom models.

Likelihood to take an action Recommendations Predicted revenue, margin, LTV Send time optimization Lead scoring Custom ML models

Likelihood to take an action

Run pre-built models to evaluate probability that a user will complete specific actions, such as making a purchase or unsubscribing from an email list.

Predictive analysis assists in identifying the probability of user actions, enabling more focused customer retention strategies. By analyzing user interaction data, you can predict the likelihood of a user taking an action.

Examples: likelihood to make a purchase, churn, watch more than 50% of a video, finish a course, read an article to the end, close a page in the next 10 seconds.

Recommendations

Provide personalized product and service recommendations based on user behavior.

Run models for product recommender and offer selector to enhance customer experience with real-time, tailored suggestions:

suggested products based on browsing history or cart contents
smart upsell recommendations
offer selection to maximize conversion rates
adaptive content delivery

Predicted revenue, margin, LTV

Leverage predictive modeling to forecast revenue, calculate margins and estimate the lifetime value of customers.

Optimize marketing costs by identifying the most valuable customer segments and tailor your budget to retain them. Focusing your advertising budget on these customers helps maximize return on ad spend (ROAS) and ROI.

Return and Refund Predictions

Determine the likelihood of returns and refunds with machine learning by incorporating factors such as purchase history and customer feedback.

The return & refund model helps identify the probability of return for customers, products or product categories.

Send time optimization

STO enables you to personalize message send time for each customer, leading to increased open rates and engagement.

Predictive send-time analysis ensures communication is received when the user is most likely to open the message.

This model can be used with email services, notifications, ecommerce services and advertising platforms to maximize engagement.

Lead Scoring and Prioritization

Score and prioritize leads based on their likelihood to convert.

ML-driven dynamic scoring helps sales teams focus their efforts on the most promising prospects:

Rank leads based on engagement, demographic data and past interactions
Allocate resources to nurture leads with highest potential

Custom ML models

Run your own models on Almeta ML with first class infrastructure, robust integrations, real-time event processing and programmable triggers.

Almeta ML supports running ONNX models on the server and in the browser.

Act on it

Use ML insights, scores and metrics to maximize ROAS and minimize churn.

Advertising

Create ML-based audiences on Google Ads, Facebook Ads, Bing Ads, TikTok Ads, and other networks.

Optimize your budget: predictive metrics help focus ad spend on users most likely to convert, leading to higher returns on ad spend.

Email targeting

Personalize email campaigns using likelihood to make a purchase, interest in a product category, probability to unsubscribe, predicted revenue.

Choose the best send time for each customer with send time optimization (STO).

You can use built-in integrations with email service providers or export lists for manual processing.

Offer selection

Use behavior predictions to optimize offer prioritization and generate sequences of personalized offers for each customer.

ML-driven insights help marketers automate their next best action and next best offer strategies.

Personalization

Personalize user experiences based on real-time behavior predictions.

Grow your sales: product recommendations, personalized search, upsells and content lead to increased conversion rates and AOV.

Retention

Identify customers that are likely to churn, and take proactive action.

Connect with customers that are likely to move to a different loyalty tier. Reward your most loyal customers to keep them active. Win back inactive customers with personalized offers and discounts.

Custom workflows

Integrate Almeta ML with your existing data platforms, introducing machine learning to your processes and enabling new workflows.

Almeta ML is launching in May 2024.
Sign up to lock in intro pricing and request early access.

No credit card required

How it works

Almeta ML is the easiest way to run machine learning calculations on your website. Most behavior-driven metrics require event tracking, which can be set up using the web tag or pre-built integrations. After that it only takes a few clicks to run as many ML models as needed.

Setup event tracking

Install the web tag or use built-in integrations.

Choose ML models

Use built-in models or upload your own. Run as many as you need concurrently.

Select destinations

Choose where you want to send the results.

While Almeta ML is completely transparent for the user, a lot is happening behind the scenes.

Events can be imported using a web tag, integrations, API, or data import. The ML service runs the calculations and sends results to selected destinations, like Facebook Ads audiences or Shopify customer profiles.

Models can trigger actions on the website with callbacks and JavaScript events.

Almeta ML can even run calculations directly in the browser: the web tag loads and runs the model for you, enabling real-time metrics for users — like probability to close the page in the next 10 seconds.

Built for developers

Machine Learning on the server and in browser

Calculate propensity to purchase or churn, product recommendations, best time to contact and other metrics with built-in models, or run any custom ONNX model.

Capture events and get calculations using a web tag, API or integrations. Run ML models on the server or on frontend.

almeta.t('event', {
    name: 'product_view',
    product_id: 12345
})

const likelihood_to_convert = await almeta.runML(MODEL_ID)
if (likelihood_to_convert > 0.8)
{
    // ...
}

Integrations

Use ML calculations with advertising platforms, email providers, eCommerce services, CDPs.

Tell us what you need

Infrastructure

Real-time event processing, programmable triggers, pre-built ML models, support for custom models.

Easy setup

Use GTM to add a web tag or use pre-built integrations. Choose the models to run and their destinations. Click “Deploy”.

Web tag and API

Utilize API or a lightweight web tag. Extend with webhooks, Lambda functions and Cloudflare Workers.

Async web tag, 3 KB gzipped, CDN.

No vendor lock-in

Export your data at any moment. Automate exports via integrations or webhooks.

Enterprise features

SSO with SAML, advanced permissions, options for hosted and on-premise.

Privacy

Almeta ML does not track any user information by default. Choose which events you want to track and which customer data to store.

No cookies by default

You can use the default setting of no cookies or track user ids from your current platform.

No personal information

No customer data is used for ML — unless you choose to include it.

Pricing

Almeta ML offers usage-based pricing with a free tier. You pay only for what you use, depending on how many events you want to track and how many model calculations you want to run. Events are pieces of customer behavior data used in your models. The number of machine learning calculations is the number of times you run your models to get calculation results.

You can sign up for a monthly plan with no commitment, or save 30% on an annual plan.

Monthly Annually — 30% off

Free

$0 / month

5,000 model calculations

10,000 events

30 days of data storage

Running models on server

Standard

$99 / month billed annually

50,000 model calculations

1,000,000 events

60 days of data storage

Running models on server

Running models in browser

Enterprise

Volume pricing

Choice of hosted / on prem

Running models on server

Running models in browser

Expedited support

Advanced permissions

SSO with SAML support

Additional model calculations: starting at $1 for 1000

Additional events: $1 for 10,000

All included usage is on a monthly basis. The Free and Standard plans include basic email support, the Enterprise plan includes expedited support. Each plan has a number of included events and model calculations. The Free plan is limited to one active web tag (which is enough for most websites).

If you have any questions or need help setting up Almeta ML with your existing infrastructure — talk to an expert.

Almeta ML is launching in May 2024.
Sign up to lock in intro pricing and request early access.

No credit card required