Key takeaways
- Uplift needs a control group; a good result on its own proves nothing.
- Change one thing at a time and keep senders, timing and follow-ups equal.
- Compare positive replies, meetings and opportunities, not opens.
- Weigh any gain against the research time it cost.
- Repeat the test on a fresh segment before changing your whole process.
What does uplift mean in a campaign
Uplift is the difference in results caused by personalisation, measured against what would have happened without it. It is a comparison, not a single number.
Without a control group you only know how the personalised campaign did, not whether personalisation helped. A good quarter, a strong list or a new offer can all lift results on their own.
Which metrics should you compare
Compare metrics close to revenue: positive replies, meetings booked and opportunities created. Treat opens as a weak signal, because many mail apps load images automatically and count opens nobody made.
- Positive replies: responses that show interest, not out-of-office notes or unsubscribes.
- Meetings booked for each prospect contacted.
- Opportunities created from each group within a fixed window.
- Unsubscribes and negative replies, which can rise if personalisation feels intrusive.
- Time spent per prospect, so you can weigh the gain against the cost.
How do you set up a fair test
A fair test changes one thing and keeps everything else equal. Take one segment, split it at random into two groups of similar size, and send both in the same week from the same senders with the same follow-ups.
Fernway, an invented HR software company, splits a list of operations directors in two. Group A gets the segment message; group B gets the same message with a first line about each company's recent hiring. Once both groups have had time to reply, Fernway counts meetings in each and the hours spent researching group B.
How do you know the difference is real
The difference is real when it holds across enough responses and repeats in a second test. Small tests swing on a handful of replies.
Before you change your whole process, check that the gap is larger than the week-to-week variation you normally see, then rerun the test on a fresh segment. If the result flips, you have not found uplift yet.
How to measure personalisation uplift
Run one test at a time, on one segment.
- Pick one segment and one personal element to test, such as a researched first line.
- Split the segment at random into a control group and a personalised group of similar size.
- Send both versions in the same period, from the same senders, with the same follow-ups.
- Record positive replies, meetings and opportunities for each group, plus the research time spent.
- Compare the groups once responses have settled, then repeat on a new segment before rolling it out.
Where Pebbles Ai fits
- CapabilityProspecting & customer recordsKeep each test group as its own segment in the CRM (in Beta), where admins can add custom fields to record the version a contact received. The CRM has no attribution, so do the comparison in your own reporting.
- CapabilityOutreach contentBuild the control and personalised versions as separate campaigns with the same personality and number of touches, so the personal element is the only difference.
Common questions
Should you use open rates to measure personalisation?
Not on their own. Opens are unreliable because many mail apps load images automatically, so replies and meetings are better evidence.
How big should a personalisation test be?
Large enough that a few replies either way would not change the conclusion. If your list is small, run the test across several weeks or segments and combine the results.
What if personalisation costs more time than it returns?
Then personalise less deeply, or only for your best-fit tier. Uplift that needs hours per prospect may not pay off for smaller deals.
About this page. Written by the Pebbles Ai team for people who run sales, marketing and growth. Last updated .
