Here is how TPV is doing
SELECT
SUBSTR(bucket_id, 6, 7) AS month,
ROUND(SUM(IF(
product_id = 'a21b31b4-74b8-4a6e-a2a8-28b85d1f8315'
AND merchant_name != 'Billbox'
AND merchant_name != 'Simpl', txn_amount_in_paise, 0)) / (100 * 10000000), 2) AS pl_tpv,
ROUND(SUM(IF(merchant_name = 'Billbox', txn_amount_in_paise, 0)) / (100 * 10000000), 2) AS billbox_tpv,
ROUND(SUM(IF(
product_id = '541eeb1c-8212-40f7-af04-05990ff00961',
txn_amount_in_paise, 0)) / (100 * 10000000), 2) AS pi3_tpv,
ROUND(SUM(txn_amount_in_paise) / (100 * 10000000)) AS total_tpv
FROM
public.transaction_events
WHERE
event_name = 'TransactionChargedEvent'
AND bucket_id >= 'date_2024_01_01'
AND SUBSTR(bucket_id, 14, 2) <= '23'
GROUP BY 1
ORDER BY 1
Overall TPV and Pi3 TPV are already at their lifetime maximum values. In a few months, PL TPV could also achieve lifetime max. If only Billbox TPV could reach 64 cr up from current 54 cr (first 23 days), we would be at lifetime max there too.
And here is how MAU are doing
SELECT
SUBSTR(bucket_id, 6, 7) AS month,
COUNT(DISTINCT IF(
product_id = 'a21b31b4-74b8-4a6e-a2a8-28b85d1f8315'
AND merchant_name != 'Billbox'
AND merchant_name != 'Simpl', user_id, NULL)) AS pl_users,
COUNT(DISTINCT IF(merchant_name = 'Billbox', user_id, NULL)) AS billbox_users,
COUNT(DISTINCT IF(
product_id = '541eeb1c-8212-40f7-af04-05990ff00961',
user_id, NULL)) AS pi3_users,
COUNT(DISTINCT user_id) AS total_users
FROM
public.transaction_events
WHERE
event_name = 'TransactionChargedEvent'
AND bucket_id >= 'date_2024_01_01'
AND SUBSTR(bucket_id, 14, 2) <= '23'
GROUP BY 1
ORDER BY 1
Pi3 MAU is at lifetime highest, but overall MAU will take a long time to catch up.
It is a long running project which suffered several delays. Last week we deployed it to prod, starting from internal user testing to 1% to 5% to 10% to 25% ramp. Thanks Raksha for following it through. We faced several bugs when deploying to internal user testing, and we fixed all these issues promptly.
We are migrating TSFE (transaction systems frontend) repo from sapper to sveltekit. Firstly, Svelte team has deprecated sapper framework. Secondly, we wanted to upgrade node version. We use node 14 which reached end of life in 2023. Lastly, build times of TSFE were 25 mins, due to outdated versions.
We are moving to sveltekit, upgrading node, upgrading Sass. Build time is now around 3 minutes. Also moved from airbrake to sentry.
Same cycle delinquency of v6.2 has been consistently better than that of v6. E.g. DPD 30 26.81% vs 25.56% for the two families in June 2nd cycle and 27.45% vs 23.91% in July 1st cycle. This week we should be able to ramp up to 100%. That would also mean some good amount of code clean up and elimination of two running pipelines.
Ripple effect of this change will continue improving our portfolio for next several months.
Fraudsters are having victims install external APKs (saying that they need this app to pay traffic fine). These APKs read all SMS and thus allow user to compromise Simpl app, OTPs, and other apps too. Team is figuring solutions to counter this.
In a significant architectural enhancement, all known users of identity service DB, which is a part of identity service RDS, have migrated to databricks as data source. Identity service will be shut down when we are sure there are no remnant use cases.
I think that even in real time datastore (which is another DB in identity service RDS), there is possibility to move users away - e.g. circuit breakers can use databricks.
Yash and Nikita published a blog article. They write on parallels between practice of DS and practice of medicine. Thanks to both! Saravanan has already connected with me for the next one on react native upgrade journey.
Last week was Shanker's last week. We wish him all the best.
Neptune dump is now available by 10 AM everyday. Earlier I think it was able late in the evening. Thanks all, specially Krishna, for making this happen. This is saving around $700 monthly cost too.
Prateek and Harsha are working on projects to approve more users and reduce delinquency. Specifically, Prateek has nearly figured a bunch of more users to approve based on num seen days, merchants seen, and amounts in approval calls. Let's see how this project goes.
Currently only SIKA merchants can use Pi3 orchestrator. Prakash and Aayush worked over the week to enable direct pi3 merchants be able to use Pi3 merchants. This would enable direct merchants (like healthkart and muscleblaze) to be able to use cross product linking.
It is coming along fine. Team has provided an ETA of Sep 26. Contracts with other teams are closed.
Project has a deadline of Oct 31. Shreyas will provide milestones and timelines this week.
Late fine waiver logic (give a max of 2 fine waivers in 6 cycles) had broken when DDD got implemented. We have now fixed it. See the write up here.
We realized that we don't need custom attribute replication in CRM. Whichever attributes are needed can be got from other services. Thanks Suresh for doing this. I love cleanups.