Typically I've found that DISTINCT is faster than GROUP BY when SELECTing a single column, i.e. as below, but obviously that could vary:
SELECT DISTINCT
SSF.customerId
FROM #example SSF
WHERE SSF.sales IS NOT NULL
--vs.
SELECT
SSF.customerId
FROM #example SSF
WHERE SSF.sales IS NOT NULL
GROUP BY SSF.customerId
SQL DBA,SQL Server MVP(07, 08, 09) A socialist is someone who will give you the shirt off *someone else's* back.