New user's journey [Web]: Delight

This issue is WIP while the team defines their plan ahead.

## Intro

**Problem we’re solving**

* Currently we offer no guidance as to the steps new users should take in product in order to get value in their first session. We also learned in our [second round of testing](https://www.figma.com/file/EajJegqEYIYK2ZJSLuLX8X/Testing-Synthesis?node-id=0%3A1) that the learning curve for Element is steep.
* Using a list of hypotheses and rapid experiments intended to simplify the journey for new users, we will improve D1. While conducting these experiments we would collect data that helps us to understand better the new user journey and flows that need to be revised later.

**Goals**

* Understand the web user's journey when they're new to Element
* Improve the user's experience in their first session

Non goals - Problems we’re not solving

* Changing fundamental flows; We want to first understand and learn the impact of actions and communication to D1, then we can move to improving the flows if needed.

**Guiding principles**

* Move fast; Be able to undo quickly anything that isn't working
* Be clear on the goals before starting; Our ability to measure experiments is limited, by being very clear on the things we want to see we are able to call pass/fail with more confidence.

**Success Metrics**
Quantitative:

* Improve Web D1 retention
* Do not decrease D7 or R30 retention
* Increase "Start tasks" and "Task completion" rates in first session

Qualitative: TBD

---

## Discovery

**Discovery output**

* Which tasks are conducive to higher retention rates, and which aren’t?
* Which tasks do users struggle with the most and what are the suggestions for improvement?
* How do we encourage users to engage with high-retention flows?
* How might we tailor a user's first session to ensure the highest chance of success?

## Delivery

* Run UserTesting sessions over time to establish a baseline and measure progress
* Craft a list of hypothesis and assumptions that we think will improve D1
* Deliver rapid experiments that improve D1 on Web

**Risks to mitigate**
| Risk | Details | Mitigation options | Decision |
|---|---|---|---|
| Increasing "Start task" decreases D1 | We may find that pushing users into task completion reveals issues with tasks that need fixing before we see significant improvement to retention or first session success. | These would require prioritisation review discussions at milestone steps in the future. | |
| Running sound experiments is not easy | - We have no A/B test infrastructure
- The user base is potentially hostile towards the idea of changing the software based on calling home
- We lack the user numbers to quickly reach statistical significance for small changes in retention
- We don't have anyone with a stats background to ensure experiments are well designed | Don't run A/B tests, instead be clear about impact and hold ourselves accountable to that. | |

---

## The Experimentation Plan

TBD: The team needs to meet and list the assumptions and hypothesis we have

Start with the Discovery and Delivery sections, then review the Success Metrics and The Experimentation Plan. The issue is still defining hypotheses and assumptions rather than naming an implementation entry point. Done is not specified; the team must first establish the plan, experiments, and criteria for measuring first-session improvement.

Written by the indexing model from the issue text.