> For the complete documentation index, see [llms.txt](https://partner-docs.covergenius.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://partner-docs.covergenius.com/optimizations/explorations.md). # Explorations Explorations are deliberately designed to observe without optimising, so that by the time you run an Experiment, you are testing a direction you already have reason to believe in. For example, an Exploration can distribute traffic across a range of price variants (e.g. -20% to +20% of your base price) and help you explore how pricing affects demand, revenue, and attach rate. The sections below walk through each part of the Explorations and the data you can expect to see in your Portal. ***Finding Explorations in your Portal:** Log in → left-hand sidebar → Optimisations → Explorations. Select any Exploration name to open its detailed view.* ## The Explorations list All active and historical Explorations are available in your Portal. The Explorations list gives you a snapshot in one place. Each entry contains:

Field	What it shows
Exploration name	The name of the Exploration and the protection product being tested.
Status	Whether the Exploration is currently Running, Scheduled, or Closed.
Type	The dimension being varied (e.g. Price).
Start date	When the Exploration began collecting data.
Duration	How long the Exploration has been running.
End date	The date the Exploration was closed.
Related Experiments	The number of Experiments linked to this Exploration.

## Deep-dive an Exploration With any Exploration in your Portal, you’ll get the detailed scope of the Exploration: * **Product.** The Cover Genius protection product being tested, (e.g. Comprehensive Travel Protection) * **Users per day.** The approximate daily traffic volume enrolled in the Exploration. Key metrics:

Statistic	What it shows
Primary metric	The primary value being measured (e.g. GWP per order).
Secondary metric	The secondary value being measured (e.g. attach rate).
Total quotes	The cumulative number of quotes generated across all variants since the Exploration began.
Duration	How long the Exploration has been running.
Variants	The number of price variants included in the Exploration.

## Variant Allocation We aim to split traffic equally across all variants. This is intentional — equal distribution keeps results unbiased, even if it takes a little longer to see where the opportunities are. The Variant Allocation table in your Portal shows how traffic is distributed across your variants and how each variant is performing against your key metrics.

Field	What it means
Variant	The variant label. The unchanged version is labelled Control. Other variants describe how they differ from it — a numeric adjustment (e.g. a price change of −20%, −10%, +10%, +20%), a different product configuration, or an alternative feature or design being tested.
Traffic %	The share of enrolled users assigned to this variant.
Primary metric	The primary value being measured.
Secondary metric	The secondary value being measured.

Unlike Experiments, Explorations don't flag a winner. Results are directional signals only. Look across the rows for meaningful patterns in the primary and secondary metrics to identify which variant is generating the best overall outcome. *Traffic is typically allocated equally across all variants. If you notice a significant imbalance, contact your Cover Genius PGM.* ## Dimension Analysis Dimension Analysis breaks down Exploration results by segment — country, device type, booking value, or customer cohort. Use Dimension Analysis to understand whether impacts are consistent across your full user base or whether certain segments respond differently to changes (e.g. pricing). A variant that performs well overall may be driven by a single high-value segment. Equally, a variant that looks neutral in aggregate may be showing strong signals in a segment that's worth targeting. Segments can be customized. Speak to your Cover Genius PGM about the segments required for your Exploration. {% hint style="info" %} **Important:** Dimension Analysis results are directional signals only. A segment-level pattern identified through Explorations is not a confirmed causal finding. Always validate with a dedicated Experiment before acting on a segment-level signal. {% endhint %} ## Traffic and Performance Traffic and Performance gives views of how your Exploration is progressing over time. * **User sessions.** Shows the volume of users assigned to each variant over time. Use this to confirm that traffic allocation has been consistent since launch and to spot anomalies. For example, a sudden drop that might indicate a configuration issue. * **Performance.** This shows how each variant's primary metric (e.g. GWP per order) has changed over time. A pattern that holds consistently over a longer duration is more meaningful than a short-lived spike. ## Segment Recommendations: Start your next Experiment Within Explorations, there is a ranked list of recommendation sets worth turning into Experiments. A recommendation set combines a specific segment (a combination of attributes that groups similar customers together) with a proposed treatment for that segment. Recommendation sets are generated automatically and screened against a set of safety checks ([the 'Hard Gate'](/optimizations/technical-reference.md#hard-gate)) before being shown. Recommendation sets are estimates only. They have not been validated by a controlled test, and they never affect your live pricing or product configuration — only a confirmed Experiment can do that. Use the recommendations to decide which segments are worth running as a dedicated Experiment. ### Reading your recommendations Each set of recommendations shows a pooled summary across all recommended segments, followed by a detailed breakdown of each one. #### Pooled summary

Field	What it shows
Pooled projected GWP uplift	The combined estimated GWP impact across all recommendation sets, if their proposed treatments were applied, translated from each set's chosen metric. Directional guidance only — verify by running an Experiment.
Pooled chance to beat Control	The combined probability that these recommendation sets, taken together, would outperform Control.
% of traffic	The share of your total Exploration traffic that these recommendation sets represent.
Segments	The number of distinct recommendation sets included here.

{% hint style="info" %} Pooled figures aren't a simple sum or average of the individual recommendation sets below. 'Pooled projected GWP uplift' can differ from the total of each set's 'Projected GWP uplift', and 'Pooled chance to beat Control' can be higher than any individual set's chance to beat Control. This is expected. {% endhint %} #### Recommendation set table

Field	What it shows
Segment	A specific combination of attributes that groups similar customers together (e.g. trip duration range combined with a partner price range).
Proposed treatment	The specific product change recommended for this segment.
Metric (control)	The baseline value of the chosen metric before any change. This is GWP per order by default, but may be set to attach rate or conversion rate depending on your account's configuration.
Metric (treatment)	The value of the chosen metric if this segment's proposed treatment were applied, compared against the control, along with this segment's chance to beat Control (P(beat)).
Projected GWP uplift	An estimate of this recommendation set's annual GWP opportunity, translated from the chosen metric. Directional guidance only — confirm with an Experiment before acting on it.
Bookings	The number of orders recorded for this segment during the Exploration period.

### Dismissing a recommendation set If a recommendation set isn't relevant to your business, you can dismiss it. Dismissed recommendation sets won't be shown again and won't be acted on. {% hint style="info" %} Dismissal availability may vary. Currently, you can reach out to your Cover Genius PGM to action a dismissal on your behalf if it isn't available directly in your Partner Portal. {% endhint %} ### Turning a recommendation into an Experiment Recommendation sets are not automatically converted into Experiments. Contact your Cover Genius PGM to set one up based on a recommendation set you want to validate. ## Related Experiments Any Experiments derived from an Exploration are captured. Each entry includes the experiment name, status, and a link to the full experiment detail. This is the feedback loop in action. An Exploration surfaces a promising direction. An Experiment validates it. The confirmed result becomes the new baseline, and the next Exploration builds on that. Over time, each cycle moves you further from intuition and closer to a continuously improving product. Use Related Experiments in your Portal to track which hypotheses from this Exploration are currently being tested. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://partner-docs.covergenius.com/optimizations/explorations.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.