Frontier 5 experiments running this week
Your AI team runs experiments on your business every week.
Different message openers, different prices, different times of day. The winners get rolled out automatically. The losers get killed. Below: 5 experiments running this week, with the current lead in each.
Experiments running this week
5
3 from sales, 2 from ops
Experiments concluded this month
12
9 rolled out, 3 rolled back
Avg lift on winning variants
+18%
trailing 30 days
WhatsApp opener test
Asking 'where to and where from' first will book quotes faster than asking for a name first, because customers want to know the move is feasible before sharing personal info.
Running Variant A (control)
"Hi! I'm the Acme Movers assistant. What's your name?"
Sample187 sessions Result22% quoted Liftbaseline
Variant B Leading
"Hi! Quick question, where are you moving from and to?"
Sample192 sessions Result34% quoted Lift+12pt
Variant B leading (+55% relative lift on quote rate)
Mid-day pricing test
Customers who get quoted between 11am and 2pm will convert at higher rates because they are researching during lunch, not after-hours when comparing.
Concluded Variant A (control) Leading
Quote on first ask at any time of day
Sample241 quotes Result$1,180 avg ticket Liftbaseline
Variant B
Quote with 'lunch-rate' nudge 11am to 2pm
Sample218 quotes Result+8% close, lower ticket Lift-3%
Control wins. Hypothesis falsified.
Weekend availability messaging
Telling customers we work weekends in the first message will increase weekend-move requests, because they don't think to ask.
Running Variant A (control)
No mention of weekend availability
Sample94 chats Result47% requested weekday Liftbaseline
Variant B Leading
'We work weekends too' in the 2nd message
Sample101 chats Result62% requested weekend Lift+15pt
Variant B leading. Weekend jobs are higher margin.
Deposit-required A/B
Requiring a $250 deposit upfront will filter out tire-kickers and increase close rate, even if it scares some away.
Running Variant A (control)
No deposit ask. Soft-book.
Sample162 bookings Result28% show up Liftbaseline
Variant B Leading
$250 deposit ask inside chat
Sample154 bookings Result72% show up Lift+44pt
Variant B leading. Net +44pt close despite 18% drop-off at deposit ask.
Follow-up cadence test
Following up 3 times in 7 days outperforms 5 times in 7 days, because over-follow-up creates ghost-mode.
Running Variant A (control)
5 follow-ups over 7 days
Sample203 cold leads Result14% reactivation Liftbaseline
Variant B Leading
3 follow-ups on day 1, 3, 7
Sample198 cold leads Result21% reactivation Lift+7pt
Variant B leading. +50% relative lift with less spam.