


Varioqub report: experiment variants and their conversion
Varioqub runs experiments on top of a Yandex Metrica counter, and Reportkey brings their results into the client report: visits, goals, conversion rate and revenue for every variant, next to advertising and overall site statistics.
- Runs on an existing Yandex Metrica connection with no extra account.
- Visits, goals, conversion rate and revenue for every experiment variant.
- Experiments in one report with Yandex Direct and site statistics, plus AI insights (not on every plan).


The Varioqub integration works through the Yandex Metrica connection: sign in with Yandex on the integrations page, pick the counter Varioqub is attached to in a project and set the variant names. Visits, goal conversions, conversion rate and revenue per variant reach the client on schedule.
Varioqub in a recurring client report
Varioqub is the A/B testing service from Yandex. It splits site visitors into groups, shows them different versions of a page and tracks which version leads to a goal more often. The service answers “which variant won”, but an agency client needs more: the link between a test and advertising, revenue per variant and a clear summary without logging into another interface. Reportkey takes experiment data and turns it into a recurring report section next to the rest of the site statistics.
How Varioqub relates to Yandex Metrica
Varioqub does not stand alone. An experiment is attached to a Yandex Metrica counter, and every participant visit gets a variant label. In the extended version of Varioqub that label is stored in raw Metrica visits alongside source, goals, device and purchases. This is how Reportkey gets the data: from the raw visit export in Metrica, not from a separate account. No additional login is needed, only the Metrica connection most clients already have.
Setting up the experiment report
If Metrica is not connected yet, sign in with Yandex once on the integrations page. In the project, pick the counter Varioqub is attached to. Then set the variant names: Varioqub stores variants as numeric IDs, and the mapping list is exported from its experiments section. Without names the report shows variants by number, which is awkward for a client.
Visits and users by variant
The first block checks whether traffic split evenly. For each variant it shows visits, users and bounces by day. If the test variant receives noticeably less traffic than planned, this is visible before anyone draws conclusions about conversion.
Goal conversion by experiment variant
The main block lists completions of the selected Metrica goals and conversion rate per variant. The goals are the ones configured in the counter: a lead, a call, an add to cart. The report shows absolute numbers and percentages, so the gap between variants reads without a calculator.
Revenue and average order value by variant
When Metrica e-commerce is set up on the site, the report adds purchases, revenue and average order value per variant. For stores this often matters more than conversion: a variant may bring more orders of a smaller size. Revenue here is what the site sent to Metrica at checkout, not a confirmed payment.
Variant traffic by source and device
The last block splits visits and conversion of each variant by traffic source and device type. It guards against false conclusions: if one variant happened to get more mobile ad traffic, the result is explained by the audience, not the design.
How the report differs from the Varioqub interface
Reportkey calculates metrics from raw visits and does not replicate the Varioqub statistical method. Significance, win probability and confidence intervals stay in the service interface. The report serves regular communication: what is being tested, how much traffic and how many goals each variant collected, and how that relates to advertising in Yandex Direct.
For mobile apps a similar role is played by AppMetrica, and for international traffic by Google Analytics 4. All web analytics tools are listed under analytics. See a sample document in the in-house demo report and terms on the pricing page.
What goes into the Varioqub report
These are the metrics that the Varioqub integration already brings into Reportkey reports.
Varioqub reporting: plans and free trial
14 days free, no card required. Prices and plan limits are on the pricing page.
Why teams connect Varioqub to Reportkey
Real reporting jobs, typical pain points, and the outcomes teams get from Varioqub in client reports.
- A Varioqub experiment is in its third week, and every Monday the client asks for a screenshot and an explanation of which variant is ahead.
- Test results live in Varioqub and ad spend in Yandex Direct, so linking a conversion lift to the budget happens by hand in a slide deck.
- A new checkout test should be judged by revenue, not only by button clicks, yet nobody calculates revenue per variant.
The team changes a lead form and tests the hypothesis in Varioqub. The weekly report shows visits and goal completions for control and test next to ad spend, so the client sees what the work delivers.
Two versions of the delivery block run on product pages. The monthly report compares purchases, revenue and average order value per variant, since Metrica e-commerce is set up on the site.
Headline and subscription price tests run in parallel. The report splits each variant's traffic by source and device, so an audience skew does not distort the conclusion.
How Varioqub appears inside the report
The report keeps the familiar Reportkey layout: source sections and period comparison.

Two columns, control and test variant, with visits, completions of the chosen goal and conversion rate, a daily conversion line for both variants below, and a table of revenue and average order value.
Three steps to the first report
Connect the account once, choose the reporting scope, and save the setup for recurring client reports.
Frequently asked questions
Short answers about how the integration works in Reportkey today.
What is required for a Varioqub report?
Does the report show statistical significance and win probability?
Why do report figures differ slightly from Varioqub?
How often is experiment data refreshed?
From which date is experiment data available?
Often connected together
Working sources that teams usually add to the same report.
Varioqub report: experiment variants and their conversion
Sign up, sign in with Yandex on the integrations page and choose the counter that runs Varioqub experiments. Your first report in 5 minutes.


