Please help with Abandoned Basket Timeline

[dreamingdigital]

Can somebody please confirm exactly what an "abandoned basket" actually is?
It is after the store settings basket timeout has been reached?
Is it after the store settings basket timeout has been reached AND the cookie has expired?
Is it when the browser window is closed or the user leaves the site?

[Brennan]

An abandoned basket is an open shopping cart that has not had any activity for a set period of time. This means that the customer has stuff in their cart but is no longer clicking around on the website, and likely has left. Miva keeps track of the last time each shopper loaded a page and we use that value to determine how long a basket has been inactive.

This is the same value used to determine the basket timeout. The one thing you want to keep in mind is that if you are deleting expired carts (which you should be) via a scheduled task, you want to make sure the Abandoned Cart email runs before the expired cart can be deleted.

The default setting are the basket becomes expired after 60 minutes of no activity and the abandoned cart email gets sent after 30 minutes of no activity.

[dreamingdigital]

Brennan
What do you recommend for web sites that have their shopping cart timeout set to one week? Or will Miva know (assuming there is other traffic on the site) that that individual customer has stopped browsing and send the 30min abandoned basket email 30min after it sees they are not there anymore? Or is is one week plus 30min. We are having a real hard time getting these to work.

[dreamingdigital]

Please respond. This is important to get sorted.

[ids]

Colin,

Here is a what I understand. For the aspect of site activity, Merchant doesn't track individual sessions for purposes of running a scheduled task. The scheduled tasks functions will trigger when there is even a single http request on merchant.mvc. I know this is the case per David H (VP). IOW, on a dev site I configured a cron job on merchant.mvc to run frequently, making sure I had a predictable Abandoned Basket Email response so I could test while building the email template functions for the Abandoned Basket campaign.

So, when the Abandoned Basket Task runs, it is triggered like every other scheduled task -- per a valid http request from any source (I imagine) the ABT will cycle through the Basket table and any qualifying basket will sent the appropriate email.

Brennan, Here's the question I want to verify. IF on the AB email (or the last one is a series) the keep-alive time is 10080 (7 days) will the Delete Expired Basket functions NOT delete these baskets being kept alive? I am fairly sure that would be the case. It's still an important question regarding the campaign I am setting up for the client. If these baskets are not available after the last AB email for a store owner specified amount of time, I am needing to develop another solution.

Thanks,

Scott

[Brennan]

That is correct. If a abandoned basket email keep alive is set to 7 days, when it runs it actually sets the last updated value for the basket 7 days in the future. So whenever the delete expired basket task runs it skips that basket since it won't be expired for 7 days in the future.

[ids]

Thanks Brennan. That does answer my question.

Scott

[dreamingdigital]

So here is a timeline scenario, please confirm I'm correct on this.
Let's say in this scenario there is continuous traffic to the site every minute of every day (since traffic is required to trigger the tasks)

Miva Settings Here:

• Store Settings Basket Timeout set to 10,080 min (7 days)
• Domain Settings Cookie is set to 525,600min (365 days - sys default) *not that this matters, right?
• First abandoned email set at 60min. (1 hr) Keep-alive set to additional 1440min
  • Does this matter? The cart setting is 7 days?
• Second abandoned email set at 24 hours. Keep-alive set to additional 1,500min (1 day & 1hr)

Real-life scenario:

1. 12:00pm > Shopper Logs into their account. (Miva now has en email address for this cart)
2. 12:00pm > Shopper adds one product to their cart but gets called out for surprise lunch meeting. Shopper locks their PC.
   1. Does it matter if the shopper leaves their browser open?
3. 1:00pm > Shopper is still out for lunch. Miva sends automated abandoned basket, right?
4. Shopper ignores the first email since they are working all afternoon.
5. Day 2 1:00pm second email gets sent out, this time store owner includes an auto-generated coupon. Shopper takes 2min and completes order.
6. Abandoned basket is now converted to order and all timers for this basket stop.

[Brennan]

All that is correct. It does not matter if the customer has the browser open or closed. We only look at the last time they interacted with the website.

[ids]

I have a campaign integrated into production that is having similar results(poor) as Colin and not sure why. My testing with short length timers on a dev site had success.

Del Exp Bask -- 1 per day in Scheduled Task
Basket Timeout is 7 days -- shouldn't matter
Keep-Alive for all 3 AB Emails is 7 days -- Doesn't affect the timers?

1st email -- Abandoned time is 30 min -- less than 7 days TO.
2nd email -- timer is 210 min -- > that the 1st AB timer
3rd email -- timer is 1440 -- > than the prev timers

Per the Scheduled Task log of the ABE it is running on schedule -- nearly to the second. Traffic on this site is not an issue at all.

The result isn't as expected.

The 1st email is being created and sent on time.
The second email -- only one of the dozens of abandoned baskets was sent. It was sent about 8 hours late instead of 3.5-4 hours later. Took 2x longer than expected
The 3rd -- no reports of a third email yet.

The first question is what could be causing the 2nd email to not trigger on time when it does?

Is there a log of which AB didn't qualify to be sent?

There is only one store setting that makes the math work for the single ABE sent (4 hrs late). The Product list Cache TO is 4 hours. Seems unlikely that number would not affect the ABE execution....right?

ATM, the only change made to try and improve the performance is the ABE scheduled task frequency went from 30 min to 15 minutes.

My thought on the series of emails, should the keep-alive time be set to the length of time you want the basket to not be expired before the next AB email will be triggered?

IOW:
1st -- 30 min, KA=210 min
2nd -- 200 min KA=1440 min
3rd -- 1400 min KA=10080 min

Thanks,

Scott

[ids]

For posterity: figured this out. The calculation is:

LastUpdated(time=0) + AB1 Abandoned timer
LastUpdated (becomes 0) + AB1 Keepk-Alive timer + AB2 Abandoned timer
LastUpdated (becomes 0) + AB2 Keepk-Alive timer + AB3 Abandoned timer

The conclusion is your desired time from 1st eligible ABE to the Next ABE is the SUM.

Scott

[dreamingdigital]

ids
Believe it or not I'm still trying to get this to work as expected. I see your calculation and I'd like to put some numbers in it. Please advise.
Say I have this setup nice and short so I can actually test this afternoon. My test site has 3 emails that get sent.
Email one after 15 min, keepalive for 2 days
Email two after 25 min, keepalive for 1 day
Email three after 35 min, keepalive for 1 day

So using your calculation I understand this:
Dec 1 @ 1:00pm shopper abandons
Dec 1 @ 1:15pm Email one sends
Dec 3 @ 1:40pm Email two sends
Dec 4 @ 2:15pm Email three sends?

[ids]

Appears you're looking for the # of minutes for the keep-alive.

Email one after 15 min, keepalive for 2 days == abandoned: 15 min, keep alive: 2880 min
Email two after 25 min, keepalive for 1 day == abandoned: 25 min, keep alive: 1440 min
Email three after 35 min, keepalive for 1 day == abandoned: 35 min, keep alive: 1440 min*

Seems your scheduled task needs to run frequently for accuracy. Maybe every 5 minutes.
* == Keep alive for the last email only effects the timing that basket will actually expire if not purchased -- which by your scenario would be 1 day after the 3rd email or 4 days 1 hr 15 min from the initial abandonment.

Scott

[dreamingdigital]

Sorry, I should have put the minutes in there. Thanks though. I mean, others reading this should have those real values.

What do you think about my timeline though? Is that correct? Dec 1, 3 and 4? Because that's not how I thought these worked and is probably a reason why testing has been so difficult. I assumed that in my scenario the entire thing would be over in 15+25+35min not 4 days, 1hr 15min. This is really good to know and have documented (again I guess since it is but I didn't get it https://docs.miva.com/how-to-guides/...-basket-emails ... did they just update that?!)

[ids]

I've tested my scenarios on a dev site with short timers. That is still not without its challenges to figure out what is really happening. But I was able to learn from the trial and error.

I think any scenario/campaign will have its own merits. IOW, tough to say what is right or wrong, good or bad. Depends on the product or service and your audience. I know -- obvious right? But how would you really know what a specific customer decides you're spamming them. That said, 15 minutes after the first abandonment seems very fast. The following emails seem ok.

BTW: it took us/me multiple passes in the documentation to grasp how the numbers work for ABE. At least we think we grasp it.

Scott