# Listing & Deal lifecycle - 2024-07-29

→ Link to the lifecycle schema: [Lifecycle states](Listing%20&%20Deal%20lifecycle%20-%202024-07-29%204dc0311b21ca44f8859969e419872ebd.md) 

This page aims to summarize the first steps conducted towards enabling a proper definition of the lifecycle of a listing and a deal, understanding deal as the unique identifier of a Host/PM/etc of our B2B clients. 

Table of contents:

The following sections focus on the lifecycle of a listing, but exactly the same logic applies to the deal lifecycle.

# Introduction

A listing, or accommodation, is the physical place where the guests accommodate into whenever they have booked it in a given timeframe from a host. Therefore, a listing corresponds to a single host, but can accommodate multiple guests over the time.

The volume of listings that we are working with provides valuable information regarding the scalability and health of our business: more listings, more bookings, more revenue.

In general though the abovementioned hypothesis might not be true since we can have some listings that are not activated by the host (meaning, these cannot accommodate more bookings). However, even if a listing is activated, it doesn’t mean that a listing is being booked - it just means it CAN be booked, but not that IS being booked. 

During the exercise of business KPIs definition, this subject has grown in attention since there’s different ways to account for the activity of a listing and, in essence, we’re interested on the one that enables us to somehow understand the potential that our business has. Even if we have some activity in the listings because these have orders, the recency and frequency of how much are they being booked could be quite interesting for our knowledge. 

Additionally, we need to take into account that eventually, a listing could churn. This means that, over a certain period of time, we would consider that this listing is inactive because it didn’t have any booking for a certain time. We could compensate those by onboarding new hosts and thus acquiring new listings, and even aiming to reactivate already churned listings. Now we’re starting to see the importance of measuring the lifecycle of a listing!

# Reasoning

In the previous section we have already identified some potential states that could help understanding in which point in the lifecycle a listing currently is on. Before jumping into the different categorisations, it’s important to understand the basis in which the lifecycle will rely on. We can consider these as ‘assumptions’, but rather more broadly as ‘reasoning’ behind the categorisation:

- We will measure the listing activity based on the bookings created in a given timeframe. At this moment, we’re not excluding any booking state - this meaning that a cancelled booking would still be taken into account to measure listing activity.
- Based on this logic, we can identify at a macro level 3 main states:
    - A listing is somehow ‘active’, in the sense that has had recently at least 1 booking;
    - A listing is somehow ‘inactive’, in the sense that has not had any booking recently, or any booking at all in its history;
    - and finally, a listing can be neither ‘active’ nor ‘inactive’, this meaning mainly that the listing has been created recently but has the potential to allocate for new bookings in the future
- Based on this logic, it makes sense to separate the path of a ‘new’ listing from those that ‘not new’, since the business strategy in these 2 areas will probably differ.
- In essence, it can be interesting to identify those natural movements between ‘new’ to ‘active’, ‘active’ to ‘inactive’ and, yes, ‘inactive’ to ‘active’.

# Lifecycle states

Based on the previous reasoning and without aiming to have a fully-detailed lifecycle, the Data team has proposed a first approach that would enable to categorise the listings per certain lifecycle states that could enable a better comprehension of the listings evolution.

This first categorisation, developed during end of Q2 2024, consists of a set of 7 mutually exclusive lifecycle states:

1. **New**: Listings that have been created in the current month, without bookings.
2. **Never Booked**: Listings that have been created before the current month, without bookings.
3. **First Time Booked**: Listings that have been booked for the first time in the current month.
4. **Active**: Listings that have booking activity in the past 12 months (that are not FTB nor reactivated).
5. **Churning**: Listings that are becoming inactive because of lack of bookings in the past 12 months.
6. **Inactive**: Listings that have not had a booking for more than 12 months.
7. **Reactivated**: Listings that have had a booking in the current month that were inactive or churning before. After the 2nd booking during the reactivation month, will be categorised as Active directly.

Below you can see a high-level schema of how the lifecycle would look like:

![Lifecycle stages with the natural transitions (continuous lines) and potential reactivation transitions (dashed lines)](Untitled%202.png)

Lifecycle stages with the natural transitions (continuous lines) and potential reactivation transitions (dashed lines)

Let’s put ourselves in the feet of a host:

Let’s imagine as a host that I add my first listing in Superhog. During the first days, it will be categorised as a New listing. If during the same month I have a booking created for that listing, then it will automatically transition to First Time Booked. On the contrary, maybe I don’t have any booking on the first month, so my listing will be categorised as Never Booked until this first activation happens. 

Once my booking has had its first booking, I’ll be categorised as First Time Booked for that given month. With or without new bookings arriving, it will automatically transition to Active on the following month. As long as I have had a booking created for the past 12 months, the listing will be considered as Active.

In the moment it’s been exactly 12 months without a booking, my listing will move towards a temporary state of Churning for a month. If no new booking is created, my listing will go towards Inactive, and it will stay here until a new booking is created - if this ever happens.

In the scenario of having a booking created while under Churning or Inactive, my Listing will activate the reactivation flow, thus moving towards Reactivated, before being considered as Active again. Here 2 things can happen: either one month has passed since I’ve been in the Reactivated stated or I’ve had more than 1 booking on the reactivation month, thus immediately moving towards Active.

Finally, the potential terminating states are the Inactive and Never Booked. Both of them can be activated or reactivated, but for most of the cases these states will act as a cemetery of listings.  

# Activity measurement

At this stage is worth noticing that some of the previous identified states indicate certain booking activity: Active - of course - but also Reactivated and First Time Booked. This is why, independently of a listing being tagged as any of these 3 states, we can go deeper on the recency of the booking to better anticipate listing churn.

We’ve also added 3 flags:

- **Has the listing been booked in 1 month?**: If a listing has had a booking created in the current month
- **Has the listing been booked in 6 months?**: If a listing has had a booking created in the past 6 months
- **Has the listing been booked in 12 months?**: If a listing has had a booking created in the past 12 months

Note that if a listing has had a booking created this month, all 3 flags will be true. Similarly, if the last booking created to a listing was 5 months ago, only the flag has_been_booked_in_1_month will be false; while the other 2 will be true. This is specially helpful to further categorise the listings that are in the Active state, since the 3 levels will apply. For the Reactivated and First Time Booked, since these states are framed whenever a booking occurs in the current month, they will always have the 3 flags as true.

This information will help categorising with different time windows the recency of the last booking created, to anticipate potential movements towards the path of inactivity.

# Deal lifecycle

Similarly as the listing lifecycle presented so far, it’s easier to identify that the same stages can apply for a deal. Deal is preferable to host or PM since it’s possible that the same deal is linked to multiple hosts. Thus, deal represents the unique B2B entity that we can consider as ‘our clients’. 

Therefore, for the Deal lifecycle we keep the same categorisation as for Listings. It has been developed during end of Q2 2024, consists of a set of 7 mutually exclusive lifecycle states:

1. **New**: Deals that have been created in the current month, without bookings.
2. **Never Booked**: Deals that have been created before the current month, without bookings.
3. **First Time Booked**: Deals that have been booked for the first time in the current month.
4. **Active**: Deals that have booking activity in the past 12 months (that are not FTB nor reactivated).
5. **Churning**: Deals that are becoming inactive because of lack of bookings in the past 12 months.
6. **Inactive**: Deals that have not had a booking for more than 12 months.
7. **Reactivated**: Deals that have had a booking in the current month that were inactive or churning before. After the 2nd booking during the reactivation month, will be categorised as Active directly.

We’ve also added the same 3 flags:

- **Has the deal been booked in 1 month?**: If a deal has had a booking created in the current month
- **Has the deal been booked in 6 months?**: If a deal has had a booking created in the past 6 months
- **Has the deal been booked in 12 months?**: If a deal has had a booking created in the past 12 months

# How does it look like?

We’ve explored in great depth the theoretical aspect of this categorisation. Now it’s time to put some numbers on the page, so we can see the volumes associated to each stage in the Listing lifecycle.

## Listing Lifecycle (as of 20th June 2024)

| Listing Lifecycle State | Has been booked in 12 months? | Has been booked in 6 months? | Has been booked in 1 month? | Volume | Share |
| --- | --- | --- | --- | --- | --- |
| 01-New | No | No | No | 4667 | 3.26% |
| 02-Never Booked | No | No | No | 95024 | 66.45% |
| 03-First Time Booked | Yes | Yes | Yes | 1526 | 1.07% |
| 04-Active | Yes | No | No | 9201 | 6.43% |
| 04-Active | Yes | Yes | No | 14991 | 10.48% |
| 04-Active | Yes | Yes | Yes | 6840 | 4.78% |
| 05-Churning | No | No | No | 947 | 0.66% |
| 06-Inactive | No | No | No | 9765 | 6.83% |
| 07-Reactivated | Yes | Yes | Yes | 32 | 0.02% |

*04-Active state total equals to 31032 listings, accounting for the **21.69%** of all listings.*

**(!) Important**: The data displayed here is taking into account the screenshot information as of 20th June 2024. Keep in mind that these values will evolve every day. 

**(!) Disclaimer**: there are some listings in the database that appear without a user host id linked, which looks suspicious. Pending checking a potential data quality issue in this area, the figures could evolve if this needs to be fixed.

## Deal Lifecycle (as of 20th June 2024)

| Deal Lifecycle State | Has been booked in 12 months? | Has been booked in 6 months? | Has been booked in 1 month? | Volume | Share |
| --- | --- | --- | --- | --- | --- |
| 01-New | No | No | No | 24 | 1.56% |
| 02-Never Booked | No | No | No | 233 | 15.12% |
| 03-First Time Booked | Yes | Yes | Yes | 52 | 3.37% |
| 04-Active | Yes | No | No | 172 | 11.16% |
| 04-Active | Yes | Yes | No | 312 | 20.25% |
| 04-Active | Yes | Yes | Yes | 605 | 39.26% |
| 05-Churning | No | No | No | 17 | 1.10% |
| 06-Inactive | No | No | No | 125 | 8.11% |
| 07-Reactivated | Yes | Yes | Yes | 1 | 0.06% |

*04-Active state total equals to 1089 deals, accounting for the 70.67**%** of all deals.*

**(!) Important**: The data displayed here is taking into account the screenshot information as of 20th June 2024. Keep in mind that these values will evolve every day. 

**(!) Disclaimer**: not all hosts have a deal associated, thus the data represented here is a lower bound estimate of the reality. Pending improvement on setting the deal information to improve data quality.