Connection note vs no note
the same request, two different bets

What each option is actually betting on

Sending a note bets that the person reads before deciding. If they do, a specific and credible line can turn a shrug into an accept, and it can also start the conversation before the connection exists. Sending no note bets that the decision happens on your profile: the photo, the headline, the shared connections, the company. That bet is cheaper, faster, and it is right more often than most people expect on genuinely cold audiences.

Notice what neither bet is about. Neither is about how good the writing is. A brilliant note to somebody who was never going to read it performs exactly as a mediocre one does, and the endless argument about opening lines is an argument about the wrong variable. The variable is whether reading is part of the decision for this audience, and that is a targeting property.

This is also why the results published elsewhere contradict each other so violently. One team runs notes on a niche of four hundred people who all know the same three names, another runs no note on a list of twelve thousand mid market managers, and both publish a rule. Both are correct about their own audience and neither has found a law.

Where a note wins

When the recipient has to place you and cannot do it from your headline. Somebody two degrees away who shares a real connection, a former colleague of a client, a person who spoke at the same event, anyone in a niche small enough that naming the right thing proves you are inside it rather than fishing at it. In those cases a single specific sentence does work that no profile optimisation will do for you.

It also wins when the sender is not an obvious peer. A junior title contacting a director, an unfamiliar company, a founder from a market the recipient does not know: all of these create a question the profile alone answers badly. A note that answers it plainly buys the accept.

One caution about what a note may not be. Invented personalisation is worse than no note at all. A line assembled from a scraped field, or a compliment about a post the sender did not read, is detectable at a glance and it costs more than the accept it was chasing, because it tells the recipient exactly what the rest of the sequence will be like. If the specific thing you would say is not genuinely specific, send nothing instead.

Where no note wins

On cold, broad audiences, and this is where the volume actually lives. When there is no shared context to reference, a note has nothing true and specific to say, so it defaults to a compressed pitch. A pitch attached to a connection request asks the recipient to evaluate an offer before they have agreed to a conversation, and that is a harder yes than the one you were asking for.

The allowance is the second reason, and it is the one people forget. Personalised notes are capped per month on a free account. If half the people you write to would have accepted on the strength of your profile, every note spent on them is one unavailable for a segment where it would have changed the answer. Treat the allowance as a budget with an opportunity cost rather than as a feature to use up.

The third reason is diagnostic. A no note programme gives you a clean read on whether your targeting and your profile are right, because there is no copy in the way. If acceptance is poor without a note, the answer is a better list or a better profile, and finding that out in week one is worth more than a note that quietly compensated for it until month three.

Who should pick which

Use a note if your addressable audience is small enough that every send matters more than throughput, if the recipients are warm or adjacent, or if you are selling into a niche where specificity is itself the credential. Under those conditions the note is not a nice touch, it is the mechanism.

Send without a note if your audience is cold and large, if your throughput is capped by the invitation ceiling rather than by list size, or if your profile already does the explaining. Most high volume programmes belong here, and most of them arrive at it after a month of writing notes that did not move the number.

Split by segment rather than choosing globally. The correct answer for your warm second degree segment and your cold industry list is usually not the same answer, and running one policy across both is how teams end up with an average that describes neither. Decide per segment, and let the note allowance follow the segment where it earns the most.

How to test it without fooling yourself

1. Split one segment, not your whole list. Mixing segments produces a result that is really a statement about which segment got which treatment.
2. Run both arms in the same week from comparable accounts. Acceptance rate drifts with account health and with the calendar, so a sequential test compares two different worlds.
3. Judge on acceptance rate first. It moves earliest and it is the number the note is supposed to affect.
4. Only then look at what happened after acceptance, because a note that lifts acceptance and depresses reply has cost you rather than helped.
5. Give it enough volume that a handful of extra accepts cannot flip the conclusion, and resist calling it after two days.

Then hold the whole programme to the downstream number rather than the argument you just settled. Our working benchmark is positive replies as a share of sends: 0.5 to 1% is workable, above 1% is strong, and under 0.5% is a signal to stop and fix targeting rather than to iterate on the opener. A note decision has never rescued a list that was wrong.