Multi-Service Users
The share of CanadaLogin users who have used it to sign in to two or more public services, out of users who have signed in to at least one.
One account for many government services is one of CanadaLogin’s main benefits. This metric measures how often that happens. It counts services, not applications, and leaves out internal CanadaLogin components.
Calculation
Read successful events from application_usage with a known user ID (not blank or "UNKNOWN"). Map each applicationid to its service through rp.alias and rp.service, and keep only services that are not internal. Drop events before the service’s launch_date, which are pre-launch testing. Count the distinct services each user has reached up to the end date, then divide users with two or more by users with at least one.
The end date is an ET day (America/Toronto). Leave out the end date to measure up to the latest event.
WITH public_aliases AS (
SELECT DISTINCT
al.alias AS applicationid,
s.service_name_en,
CAST(s.launch_date AS DATE) AS launch_date
FROM rp.alias al
JOIN rp.service s ON al.rp_id = s.rp_id
WHERE NOT s.is_internal
),
events AS (
SELECT
userid,
applicationid,
-- time is a UTC string; report on ET days
date(with_timezone(CAST(substr(time, 1, 19) AS timestamp), 'UTC') AT TIME ZONE 'America/Toronto') AS et_day
FROM ibm_verify_events_raw.application_usage
WHERE result = 'success'
AND userid IS NOT NULL
AND userid NOT IN ('', 'UNKNOWN')
),
user_services AS (
SELECT
e.userid,
count(DISTINCT p.service_name_en) AS services
FROM events e
JOIN public_aliases p ON e.applicationid = p.applicationid
-- events before launch are pre-launch testing
WHERE e.et_day >= p.launch_date
AND e.et_day <= DATE '2026-11-15'
GROUP BY e.userid
)
SELECT
count(*) AS users,
count_if(services >= 2) AS multi_service_users,
CAST(count_if(services >= 2) AS DOUBLE) / count(*) AS multi_service_share
FROM user_services;public_aliases <- tbl(con, in_schema("rp", "alias")) |>
inner_join(tbl(con, in_schema("rp", "service")), by = "rp_id") |>
filter(!is_internal) |>
distinct(applicationid = alias, service_name_en, launch_date)
tbl(con, in_schema("ibm_verify_events_raw", "application_usage")) |>
filter(
result == "success",
!is.na(userid), !userid %in% c("", "UNKNOWN")
) |>
# time is a UTC string; report on ET days
mutate(et_day = sql("date(with_timezone(CAST(substr(time, 1, 19) AS timestamp), 'UTC') AT TIME ZONE 'America/Toronto')")) |>
inner_join(public_aliases, by = "applicationid") |>
# events before launch are pre-launch testing
filter(
et_day >= as.Date(launch_date),
et_day <= as.Date("2026-11-15")
) |>
group_by(userid) |>
summarise(services = n_distinct(service_name_en), .groups = "drop") |>
summarise(
users = n(),
multi_service_users = sum(as.integer(services >= 2), na.rm = TRUE)
) |>
mutate(multi_service_share = as.numeric(multi_service_users) / users)Interpretation notes
Online services, not applications. Related applications of one service count as one service, through the registry in rp.alias. For example, MyCGC runs two applications in IBM Verify (.NET and Power Pages), and a user who reaches both counts as one service. Counting by application instead would miscount those users as multi-service.
The denominator is users who have used at least one service, not all CanadaLogin accounts. It counts users who have signed in to at least one public service. Accounts that never reached one, such as those that stopped at sign-up or only used internal CanadaLogin components, are left out. See New Accounts and Idle Account Rate for those.
Since launch, not a window. The share covers every event up to the end date, so it tends to grow as users have more time to reach a second service. To count services reached within a window instead, add a start date next to the end date.
Alternative names
- multi-service share
- users of two or more services
- cross-service users
- service breadth
- one account, many services