Rapport-first call script with more gift options: +29% donation rate
Test run byMarketingExperimentsResearch publisher
Commentary byLiftArchiveOur analysis, not theirsA rewritten call-center script lifted a nonprofit's donation rate from 14.32% to 18.47%, a 29% relative increase.
MECLABS Institute ran the test for a nonprofit research partner that raised donations by phone, and it was published on MarketingExperiments.
Three changes shipped together, and the two scripts ran at two different call centers, so this is not a clean randomized split.
- Donation rate rose from 14.32% to 18.47%, a 29% relative lift, reported at a 99% level of confidence.
- One-time gifts rose 90%, which suggests the extra cause-related ask did a lot of the work.
- The test covered 19,655 calls, a large sample for a phone channel.
- Three changes shipped together, so the lift belongs to the script as a whole, not to any one line.
Control
Agents went straight into asking for a donation without first connecting with the caller. The only amount mentioned was $20 a month.
Variant
Agents asked for the caller's name and thanked them for past giving before the ask. They suggested $20, $40 or more a month, and asked suitable callers for an extra one-time gift tied to a current cause.
Why it likely won
Our reading of the result, not the practitioner's claim.
- Opening with the caller's name and a thank-you likely made the ask feel less abrupt and the caller more open.
- Naming $40 next to $20 removed the implied ceiling of a single suggested amount.
- A one-time gift for a timely issue gave callers who could not commit to more monthly giving another way to say yes.
- The source notes donors skewed older, and a one-off gift may be an easier commitment than a bigger monthly pledge.
Caveats: The control and new script ran at two separate call centers rather than on randomly assigned calls, so differences between the centers or their agents may account for part of the lift. Three script changes were bundled, and agents were allowed small wording tweaks after a pretest.
Who ran it
MarketingExperiments ran this test for a nonprofit raising donations by phone and published the results (MarketingExperiments). The numbers are theirs; the summary and commentary on this page are ours.
Read the original case study on MarketingExperiments →
How to run this test on your site
- Hypothesis
- If we open calls by building rapport, offer a choice of monthly amounts and add a one-time gift ask, then more callers will donate because the ask feels personal and gives them more than one way to give.
- Primary metric
- Donation rate per call
- Guardrails
- Average gift size
- Call handling time
- Monthly donor retention
- Who should test it
- Nonprofits and lead-gen businesses with inbound or outbound phone teams and enough call volume to split calls randomly within one center. It is less useful if you take only a few calls a day, or if agents won't stick to a script.
Questions
Can you A/B test a call center script?
Yes. The cleanest way is to randomly assign calls or agents within the same center to each script.
Running each script at a different center, as here, is easier but mixes the script effect with differences between the teams.
Does offering more donation amounts lower the average gift?
It can go either way. Here the higher suggested option was paired with the old one, which avoids anchoring people only on the lowest amount. Track average gift size alongside donation rate.
Should a donation call ask for a one-time gift as well as a monthly one?
In this test one-time gifts rose 90% when agents asked for one.
Asking only after a monthly decision, and skipping it for callers who mention money trouble, keeps it from feeling pushy.