I run LinkedIn and email outreach for a consulting business. Around 100 prospects, fully personalised: I write every message myself upstream and import them as custom variables, one column per step. Demanding, but exactly what the product advertises. Here is what happened in two weeks.

1. The CSV import puts values in the wrong field. I imported a clean 7-column file: 85 rows, UTF-8, no line breaks, no quotes, every single cell populated. Three of the four message columns landed correctly for all 85 prospects. The fourth, followup, reached custom_followup for only 60. For the other 25, the text was imported but written to custom_email_original (18 prospects) or custom_first_name_original (7). The error is stable per prospect, not per file row, and the affected prospects are scattered through the file with nothing in common.

2. One empty cell stops the campaign for everybody. Because those 25 now had an empty custom_followup, the entire campaign stopped. Not for the 25 — for all 96.

3. Support told me my file was incomplete. It was not. Their diagnosis was that "for some contacts in the CSV file you uploaded, there is no data available for the custom variables". Every one of my 85 rows had all 7 fields populated. I can produce the file. The values were missing from the campaign because their own import had written them to a different column. I was told to fix a problem I did not create, and no one ever looked at the import itself. When I escalated with the figures above, the answer was a fallback-phrase tutorial.

4. A fallback phrase is not a fix, it is a gag. Their recommended solution is to add a fallback so the campaign doesn't stop. That means sending a generic sentence instead of the message I wrote, to the 25 prospects whose content their import misplaced. The campaign keeps running and the damage becomes invisible. That is not a correction, it is a cover-up of their own defect.

5. It displays an email address and simultaneously reports it cannot find one. On a separate email campaign, 51 prospects had their address populated in both the email and found_email columns of Closely's own export, while email_found_error was 1 and email_sent was 0. Three columns of the same row contradicting each other. Those 51 were never emailed, and most were then marked finishedNoReply — recorded as not having replied to a message that was never sent. Addresses I supplied were silently ignored. I only found out by cross-checking their export against my source file.

6. You cannot edit a running campaign. When something is wrong, you cannot fix it. You rebuild from scratch. And you cannot remove a prospect from a campaign either — only stop them.

7. The support chat frequently fails to load, which is a problem when the tool it supports breaks this often.

Verdict. The product can put your data in the wrong field, stop every campaign because of it, and tell you it was your fault. If you run simple campaigns with short variables like first name and company, you may never hit this. If you rely on per-prospect personalised content imported from a file, you will need to export every campaign and compare it column by column against your source. I had to write a script to do that. An outreach tool should not require an audit tool, and a vendor should not blame the customer's file before reading its own.