ABtasty

AB Tasty is an A/B testing and personalization platform that helps companies increase conversion and revenue through rapid experimentation, Bayesian analytics, and developer-free visual editing tools.

Who it's good for and when it's good for

The service is aimed at marketing teams, product managers and analysts who have an active website or web application and tasks to increase conversions and LTV. It is especially in demand in eCommerce, fintech, media and services with regular traffic.

• eCommerce - tests of product cards, shopping cart, checkout, promo widgets.
• Services and SaaS - onboarding, pricing and plans, enrollment forms.
• Media and content - recommendations, banners, navigation.
• Marketplaces - filters, giveaway, experimenting with listings.

Key features

AB Tasty covers the entire experimentation cycle - from creating variations to deciding on the results, taking into account audience segments and actual user behavior.

• A/B/n testing and split by URL to test different versions of pages.
• Multivariate testing for evaluating item combinations.
• Predictive testing and dynamic traffic distribution based on the principle of multi-armed bandits - speeds up getting winning variants.
• WYSIWYG editor and code editor - quick edits and custom changes.
• Unlimited variations, flexible percentage traffic allocation.
• SPA support, QA test mode, previews.
• Stat Engine based on Bayesian statistics - understandable probability of superiority of a variant.
• Real-time results, reports and dashboards.
• Multiple targets in one test - from clicks to transactions.

Features are combined in a single interface, which speeds up the launch of experiments and reduces the development burden.

Types of experiments and when to use them

Metrics, targets and segments

The service allows you to customize business goals and analyze results by segment to see the impact of changes across different audiences and devices.

• Targets for URLs, interaction targets - clicks, bounce, registration.
• Transactional targets - revenue, deposit, loan, and other metrics.
• Custom targets - events, element clicks, target action.
• Cross-domain tracking - correct attribution in complex funnels.
• Data filtering and segmentation - device, source, behavior.

Average check and key conversion step metrics are available for online stores based on all experiments.

Personalization and widgets

A dynamic library of widgets helps you quickly launch personalized scripts without development and tailor content to audience segments.

• Pop-ups, banners, notification bars.
• Polls, chat elements, prompts and promo blocks.
• Targeting rules - by behavior, device, traffic source.

Such elements can be used as an addition to the visual editor or as standalone solutions to quickly influence KPIs.

Integrations

AB Tasty works together with popular analytics systems to facilitate hypothesis testing and end-to-end impact assessment of experiments.

• Ready-made connectors with Google Analytics and Adobe Analytics.
• Export slices and compare results by segment.

How it works in practice

The implementation process is transparent and iterative: from hypotheses to decision making.

1. Hypothesis formulation based on data and user insights.
2. Create variations via WYSIWYG or code, customize audiences and traffic.
3. Startup with QA mode, correctness and stability control.
4. Real-time data collection, analysis with Bayesian estimation.
5. Decision making - scale the winner or start the next cycle.

What you get

Experimental results are organized into visual reports with key metrics so that the team can make quick decisions.

• Key event conversion and bottom line revenue.
• Average check and the impact on the shopping cart.
• Involvement and quality of interactions.
• Comparison across devices and browsers.

Restrictions and requirements

AB Tasty is aimed at companies with a team and an existing web product - there is no self-service demo version, access is available on request. For statistically significant results, sufficient traffic and time for data collection is required.

Cost

Prices and package composition are available on request and depend on the scope of the project and the set of personalization and testing modules. The platform is used by thousands of companies, including international ones, and is supported by teams in different countries.