Skip to content
  • There are no suggestions because the search field is empty.

Technical FAQ

Technical Questions when getting started with Visually

How does Visually affect website loading speed?

Visually is designed with performance and scalability in mind. The primary asset loaded directly on the website is our lightweight SDK, which is only 16KB in size.

All additional resources are loaded asynchronously and in a non-blocking manner, ensuring they do not interfere with the site’s rendering or user experience. The SDK itself is hosted on Google CDN, providing highly reliable global delivery with extremely fast response times.

Key Performance Metrics

  • SDK size: 16KB
  • Average SDK round-trip latency: ~22ms
  • Resources loaded asynchronously
  • Non-blocking architecture
  • Google CDN delivery

Our architecture is specifically optimized to minimize impact on Core Web Vitals and overall page responsiveness.

How does Visually affect Google Lighthouse scores?

Visually may introduce a slight impact on Lighthouse performance scores in some environments. Based on both internal monitoring and Shopify ecosystem observations, the impact is typically:

  • Average impact: ~1-2%
  • Worst-case scenarios: up to ~10%

However, Lighthouse scores do not always represent real-world user performance accurately, especially for highly dynamic ecommerce environments. In practice, Visually is engineered to maintain performance levels competitive with — and often better than — leading experimentation platforms such as Dynamic Yield and Google Optimize.

We continuously optimize our SDK and delivery mechanisms to minimize performance overhead while maximizing experimentation flexibility.

How does Visually maintain low latency at scale?

As the number of active experiences grows, Visually employs multiple optimization strategies to ensure latency remains low and scales linearly.

Performance Optimization Techniques

Smart Experience Targeting

Only experiences matching the visitor’s targeting rules are delivered, minimizing unnecessary payload size.

Intelligent SDK Cache Layer

The SDK caches previously fetched experiences and only requests incremental changes, reducing repeated data transfer.

Dynamic Client-Server Architecture

Instead of generating a large monolithic JS/CSS bundle, experiences are dynamically served and injected as needed.

DNS Prefetching

We minimize DNS resolution latency using prefetching techniques to accelerate network initialization.

Dynamic Widget Injection

Widgets are injected into the page only when required, reducing unnecessary rendering and execution overhead.

These optimizations allow Visually to deliver highly personalized experiences while maintaining fast and responsive website performance.

How accurate is Visually’s tracking system?

Achieving perfect attribution and tracking accuracy across all browsers and environments is inherently challenging due to factors such as browser restrictions, cookie policies, ad blockers, and cross-device behavior.

When sampling website performance and user behavior Visually typically achieves an effective accuracy rate of approximately:

  • 75%-95% effective accuracy

This accuracy level reflects real-world ecommerce environments and aligns with the limitations faced by modern analytics and experimentation platforms.

How is Visually improving tracking accuracy?

Visually is actively developing an AI-powered identity resolution system designed to improve session stitching and cross-session attribution (for consent users).

The solution leverages:

  • Behavioral event analysis
  • Client-specific fingerprinting signals
  • Session correlation models
  • AI-driven user matching

This approach enables more accurate reconstruction of user journeys across sessions and devices, resulting in improved personalization and reporting accuracy.

Can customers use custom tracking IDs or cookies?

Yes. Visually fully supports custom tracking implementations.

Customers can integrate:

  • Custom cookies
  • User IDs
  • External tracking identifiers
  • CRM/customer identifiers
  • Proprietary attribution systems

Our engineering team can assist with implementing custom tracking logic to ensure compatibility with existing analytics and data infrastructures.

What technology powers Visually widgets?

Visually widgets are built using Preact, a lightweight alternative to React.

Why Preact?

  • Extremely small footprint (~3KB)
  • Fast rendering performance
  • React-compatible architecture
  • Minimal impact on website performance

Each widget operates as an independent component dynamically injected by the Visually SDK.

This modular architecture enables:

  • Faster execution
  • Improved maintainability
  • Independent rendering
  • Flexible deployment

Is it mandatory to use Visually’s built-in widgets?

No. While Visually provides a large and continuously growing library of production-tested widgets, customers are not limited to using only prebuilt components.

Available Options

Native Visually Widgets

Ready-to-use widgets optimized for performance and conversion.

Manual or AI Generated Code Blocks

A native feature allowing customers to inject custom:

  • HTML
  • CSS
  • JavaScript

This enables teams to rapidly prototype and validate ideas before permanently implementing them into their codebase.

Feature Flagging & Native Testing

Visually can also be used to:

  • Activate feature flags
  • Test native website changes
  • Run experiments on existing frontend functionality
  • Run Shopify theme/template based testing

This provides maximum flexibility for both marketing and engineering teams.

How does Visually’s recommendation engine work?

Visually’s recommendation engine combines multiple machine learning and rule-based approaches to generate highly relevant product recommendations.

Depending on the strategy, the platform may use one or several algorithms simultaneously.

Recommendation Algorithms

Content-Based Filtering

Recommends products based on product attributes, metadata, and similarities to user preferences.

Association Rules

Identifies products frequently viewed or purchased together using historical behavioral patterns.

Collaborative Filtering

Uses similarities between users and products to generate personalized recommendations.

Includes:

  • User-based collaborative filtering
  • Item-based collaborative filtering

Deep Learning

Neural network models trained on user interactions and behavioral signals to uncover complex recommendation patterns.

Algo / Strategy

Content Based

Association Rules

Collaborative Filtering

Deep Learning

Custom Logic

Manual

✓

        

Most Popular

✓

        

New Arrivals

✓

        

Cart Items

✓

        

Recently Viewed

✓

        

Purchased Items

✓

        

Viewed Together

✓

  

✓

    

Purchased Together

✓

✓

✓

    

Personalized

✓

  

✓

✓

  

Advanced Rules

✓

      

✓

Viewed with Recently Viewed

✓

✓

      

Purchased with Recently Purchased

✓

✓

      

What is your official SLA?

Uptime/Availability: 99.5% Yearly Uptime

Response Time:

  • Critical issues: Typically within 1 hour.
  • High-priority issues: Typically within 4 hours.
  • Medium-priority issues: Typically within 8-24 hours.
  • Low-priority issues: Typically within 48 hours.

Resolution Time:

  • Critical issues: Typically within 24 hours.
  • High-priority issues: Typically within 48 hours.
  • Medium-priority issues: Typically within 3-5 business days.
  • Low-priority issues: Typically within 10 business days.

Maintenance and Updates:

  • Scheduled maintenance: Typically outside of peak hours with prior notice.
  • Emergency maintenance: Typically communicated as soon as possible.

Our status page can is updated in real time and can be accessed from: https://status.visually.io/

What gets installed when I install Visually?

Visually is installed directly from the Shopify App Store and integrates with the storefront using Shopify’s recommended Online Store 2.0 architecture.

During installation, Visually injects a lightweight snippet called: visually_io_sdk.liquid into the store’s snippets and referencing that script from the theme.liquid file, immediately after the <head> tag.

This integration enables Visually to dynamically load and manage experiences across the storefront without requiring manual deployment work from the merchant.

It’s worth noting that when uninstalling Visually you’ll need to manually remove these lines or just follow our official help center article: https://help.visually.io/how-to-manually-remove-visually-code

How does Visually work on the storefront?

Fetching Experiences & Targeting Rules

Once the Visually runtime loads, it securely communicates with Visually’s backend to determine which experiences should be applied for the current visitor and page.

These experiences can include:

  • Visual Editor changes
  • A/B test variants
  • Personalization campaigns
  • Product recommendation widgets
  • Feature flags
  • Audience targeting rules

The decision engine evaluates multiple factors in real time, such as:

  • Current URL/page
  • Device type
  • Visitor behavior
  • Audience segments
  • Experiment allocations
  • Customer attributes

Only the experiences relevant to that specific visitor are returned and applied, helping minimize unnecessary data transfer and maintain fast performance.

Non-Destructive Client-Side Modifications

Visually applies changes directly within the browser using lightweight DOM and CSS manipulations.

This allows teams to modify the storefront experience without permanently changing the Shopify theme codebase.

Supported Modifications Include

  • Updating text and images
  • Changing styles and layouts
  • Reordering or moving elements
  • Showing or hiding sections
  • Injecting UI components such as:
    • Banners
    • Popups/modals
    • Badges
    • Recommendation widgets
    • Promotional elements

Because all modifications are applied dynamically at runtime:

  • Original theme files remain untouched
  • No permanent code changes are made
  • Disabling Visually immediately restores the original storefront experience

This “non-destructive” architecture significantly reduces deployment risk and allows teams to experiment safely without affecting core storefront stability.

Routing & Dynamic Navigation Support

Modern Shopify themes increasingly use dynamic page navigation and partial page rendering instead of full page reloads.

Visually’s runtime automatically detects:

  • Route changes
  • AJAX navigation events
  • Dynamic content updates

When navigation occurs, the runtime intelligently re-evaluates and reapplies the relevant experiences for the new page state.

This ensures experiences remain consistent and functional across:

  • Collection pages
  • Product pages
  • Search/navigation flows
  • Single-page-app-like storefront behaviors

What data does Visually store?

Visually only stores the data required to power experimentation, personalization, analytics, and recommendations.

Data We Store

  1. Catalog Data

Collected via the Shopify API, including:

  • Products and variants
  • Meta-fields
  • Inventory availability
  • Images
  • Pricing
  • Collections and related catalog metadata
  1. Anonymized Order Data

Collected via the Shopify API, including:

  • Order totals
  • Taxes
  • Shipping amounts
  • Purchased items
  • Discount codes

No personally identifiable customer information is stored.

  1. Visual Analysis Data

Collected by Visually’s automated analysis agents, including:

  • Website styling and structure
  • UI components and features
  • Cart implementation type
  • Theme behavior and storefront capabilities

This data helps optimize compatibility and personalization capabilities.

  1. Anonymized Visitor Data

Collected through Visually’s tracking infrastructure, including:

  • Page views
  • Behavioral events
  • Session activity

What Visually Does NOT Store

  1. Personal User Information

Visually does not store:

  • Customer names
  • Emails
  • Phone numbers
  • Addresses
  • Payment information
  • Any personally identifiable information (PII)
  1. IP Addresses

Visually does not retain visitor IP addresses.

Consent Management & CMP Integration

Visually integrates with Shopify’s official Customer Privacy & Consent APIs and respects the merchant’s configured consent management platform (CMP) settings.

Supported Consent Modes

1. Opt-In Consent (GDPR / EU)

Used primarily for GDPR-compliant regions such as the European Union.

In this mode:

  • Visually does not initialize tracking
  • No experiences are loaded
  • No storefront modifications are applied
  • No visitor data is collected

until the visitor has explicitly opted in to tracking and personalization cookies.

This means Visually remains completely inactive until consent is granted.

2. Opt-Out Consent (California / CPRA)

Used primarily for California privacy regulations.

In this mode:

  • Visually loads by default
  • Visitors may opt out of tracking or personalization
  • Consent preferences are continuously respected and updated

3. No Consent Required

Used in regions where explicit consent is not legally required.

In this mode:

  • Visually loads immediately
  • Experiences and personalization are applied normally
  • Standard anonymous tracking operates automatically

Is Visually SOC 2, GDPR, or CPRA compliant?

Visually is currently in the process of completing formal compliance and audit programs for:

  • SOC 2
  • GDPR

Our infrastructure and platform architecture are being aligned with industry-standard security, privacy, and operational best practices as part of these initiatives.

Current Privacy & Security Practices

Even while formal audits are in progress, Visually already implements multiple privacy and security-focused practices, including:

  • Anonymous visitor tracking
  • No storage of personally identifiable information (PII)
  • No retention of IP addresses
  • Encryption in-transit and in-rest
  • HTTPS-secured communication
  • Shopify-native consent integration
  • Configurable consent management modes
  • Minimal and non-destructive storefront integration

How does Visually affect SEO?

Visually is designed to minimize SEO impact and preserve the store’s native Shopify SEO structure.

Additionally, most known search engine crawlers and SEO agents are intentionally excluded from Visually experiences and tracking logic. In many cases:

  • Visually does not initialize for recognized bots
  • Personalized experiences are not rendered for crawlers
  • SEO agents typically index the original storefront content rather than Visually modifications

This helps ensure:

  • Stable SEO indexing
  • Consistent crawl behavior
  • Reduced risk of SEO inconsistencies caused by dynamic personalization or experimentation

Combined with our lightweight, non-destructive client-side architecture, this allows merchants to safely run personalization and experimentation initiatives without significantly affecting organic search visibility.

Do you use customer data to train AI models?

No.

Visually does not use merchant or visitor data to train public or shared AI models.

Customer data remains isolated to the merchant’s own environment and is only used to power the specific features and experiences enabled for that merchant, such as:

  • Personalization
  • Recommendations
  • Experimentation
  • Analytics

We do not use customer data to improve or train foundation models for other customers or third parties

Which AI models does Visually use?

Visually currently utilizes the Google Gemini family of models for selected AI-powered capabilities and internal processing workflows.

The exact model usage may evolve over time as the platform and AI infrastructure continue to develop.

How does Visually minimize website flicker?

Visually is designed to apply experiences as early and efficiently as possible in the page lifecycle, and in most cases users will not notice any visual flicker.

Because experiences are loaded dynamically in the browser, small visual transitions can occasionally occur when modifying elements that are immediately visible above the fold during initial page render.

When can flicker occur?

Flicker is most likely to occur when:

  • Editing above-the-fold content
  • Replacing large hero sections
  • Reordering visible layout elements
  • Applying major styling changes immediately after page load

This is a common challenge across all client-side experimentation and personalization platforms.

Visually Anti-Flicker Protection

To minimize this effect, Visually provides an optional anti-flicker feature that can be enabled:

  • For specific elements
  • For individual experiences
  • Or for the entire page

The anti-flicker mechanism temporarily hides targeted content until the relevant experience is fully evaluated and rendered, preventing users from briefly seeing the original version before modifications are applied.

This helps create a smoother and more visually stable experience for sensitive or highly visible page areas.

How does Visually maintain long-term visitor identity across personalization campaigns?

Visually maintains long-term personalization consistency using anonymous visitor identifiers and behavioral session history stored within our tracking infrastructure.

Each visitor receives a Visually-generated anonymous ID that allows the platform to:

  • Recognize returning visitors
  • Persist audience memberships
  • Maintain experiment allocations
  • Apply consistent personalization across sessions
  • Build long-term behavioral profiles and affinities

How is personalization continuity maintained?

As visitors interact with the storefront, Visually continuously updates behavioral signals, these signals are associated with the anonymous visitor profile and used to:

  • Reapply relevant experiences
  • Maintain segmentation consistency
  • Improve recommendation quality over time
  • Support lifecycle and retention campaigns

By default, personalization continuity is maintained at the anonymous browser/session level. 

For advanced implementations, merchants may optionally provide:

  • Internal Customer IDs
  • Edge user ids (for example using edgemesh id)
  • Loyalty IDs
  • Custom hashed identifiers

This allows Visually to associate activity across:

  • Multiple sessions
  • Logged-in experiences
  • Different devices

while still remaining fully merchant-controlled.