How does OFN count the number of incomplete orders?

[RachL]

Description

I have a user with 747 uncompleted orders. He is a bit worried because he never had any real orders. So he is afraid people don't manage to order...

Steps to Reproduce

See uncompleted orders for hub 730 in French production

Animated Gif/Screenshot

Severity

bug-s3: a feature is broken but there is a workaround

Your Environment

• Version used: v2.9.2
• Browser name and version: Firefox 76
• Operating System and version (desktop or mobile): Desktop

[lin-d-hop]

This number is based on the number of incomplete carts.
But for some reason OFN creates WAAAAY more carts that people actually need. A shopper will create a bunch of carts in the process of getting to checkout.

This quirk is coming up in all kinds of discussions at the moment so I think we might be working on it soon :-)

[luisramos0]

👍
I looked at FR live and this user is an outlier with 700+ carts, but I saw quite a few users with around 100 carts...
We need to create a new cart when the user changes shop but I dont think we should create a new cart when the user comes back after a few days for example.

[lin-d-hop]

I want to solve this properly as UK's S3 this round

@openfoodfoundation/train-drivers-product-owners We need to define the correct behaviour in order to solve this issue :-)

[RachL]

I would say that as soon as an item is put in cart, the order should show as incompleted.

If the shopper comes back, it would be nice to use the same cart and not open a new one.

All the other cases where we open a cart (like when someone is visiting a shop), if we need them for technical reason, let's keep them. But we need to remove them from the list of uncompleted orders?

[jaycmb]

I would say that as soon as an item is put in cart, the order should show as incompleted.

Current behaviour is order is in state cart as soon as items are selected (and stays in this state until payment +shipment selected and shopper clicked 'place order').

All the other cases where we open a cart (like when someone is visiting a shop)

We open a cart automatically when someone is visiting a shop?!

If the shopper comes back, it would be nice to use the same cart and not open a new one.

That would be nice indeed, but sounds like more complicated.
Should that involve showing the shopper previously saved items in the cart? Or only "reuse" previously generated cart/order if a user returns within a certain time period after having abandoned the cart (e.,g. because switching to another shop within same session)?

[RachL]

Current behaviour is order is in state cart as soon as items are selected

@jaycmb are you sure of this? Many of the cart currently opened do not have any items.

We open a cart automatically when someone is visiting a shop?!

More specifically, each time the person is login back in, see the discussion that has generated the issue:

https://openfoodnetwork.slack.com/archives/CG7NJ966B/p1586458155387700

but sounds like more complicated.

I'm not so sure about that. It sounds like Luis mentioned it above ☝️

We need to create a new cart when the user changes shop but I dont think we should create a new cart when the user comes back after a few days for example.

If the user switch shops he is loosing the cart, I think we should keep this behavior.

[RachL]

@apricot12 @jibees if one of you picks this s3 up, the only task for now is to do a spike to understand in which cases today we open a cart. From there we will create other issues to remove or add cases where we open a cart.

Is the data in this issue useful or should I close it and just open a spike issue?

[jibees]

Regarding the code, and assuming that an order (ie. a cart) is opened when calling lib/spree/core/controller_helpers/order.rb@current_order(true), we do create an order (if it does not exist) in the following place:

• app/controllers/base_controller.rb#38: when we set an order_cycle
• app/controllers/cart_controller.rb#7: POST on /cart/populate. Called by app/assets/javascripts/darkswarm/services/cart.js.coffee@adjust() via update() each time we edit a ship variant
• app/controllers/enterprises_controller.rb#66: GET on /:id/shop when order has insufficient_stock_lines (and then redirect to /)
• app/controllers/enterprises_controller.rb#70: when calling reset_order ie. each prepend_before_action on /shop
• app/controllers/shop_controller.rb#15: POST on /shop/order_cycle
• app/controllers/spree/orders_controller.rb#43: GET on /orders/:id/edit
• app/controllers/spree/orders_controller.rb#112: PATCH on /orders/:id
• app/helpers/shared_helper.rb#5: called via view app/views/shared/_distributor.html.haml
• app/services/order_completion_reset.rb#28: seems to be used for the "continue shopping" feature once an order is completed.
• lib/spree/core/controller_helpers/order.rb#80: not sure to understand but seems to be called each time we use Order (via before_action on base.class_eval)...

Saying that, I'm afraid that it is totally useless, because it seems to be called/used a lot.

(Next step: try to recreate each step functionally (ie. as a user/admin) that leads to these lines.)

I'm wondering if there is an opportunity to log (with a certain level of severity, or a certain tag, or whatever, ...) each time an order is actually created (on lib/spree/core/controller_helpers/order.rb#41) with its own stack and then understand where does this empty cart comes from.
Certainly a question more for @openfoodfoundation/core-devs

[Matt-Yorkley]

Yeah, there's lots of places where we create or discard carts for various reasons. I'm not sure we can easily remove them...

[kirstenalarsen]

This is really reassuring to me - so all those incomplete orders are not actually abandoned carts. It would be really good to have an understanding of how many of them are (and a way to track that) as that’s a metric I’d like us to be actually focusing on

…

On 3 Sep 2021, at 6:49 am, Matt-Yorkley ***@***.***> wrote: Yeah, there's lots of places where we create or discard carts for various reasons. I'm not sure we can easily remove them... — You are receiving this because you are on a team that was mentioned. Reply to this email directly, view it on GitHub <#5236 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAWGXSH47CWV6DGYOSBWKBTT77PNLANCNFSM4MIYTETA>. Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[mkllnk]

I guess that you need to look at carts with items in there and ignore the empty carts because they get created all the time.

[RachL]

thanks @mkllnk for the suggestion! At delivery train today we decided that as a first step here we would hide uncompleted order with no line items from the user. This should close this s3 for now. @jibees it's all yours!

In parallel, we need to work on our metrics like conversion rates, both for our users but also to correctly assess the impact of the new slipt checkout work.