A server drops the check at a table, and with it the offer of a piece of chocolate for every guest. Then the interesting part. She takes a step away, stops, turns back, and offers a second piece as a spontaneous extra for those particular people. The tables in that condition tipped an average of 22.99 percent of the bill. Tables that got the same two chocolates offered up front, as plain policy, averaged 21.62 percent. Tables that got no candy at all averaged 18.95 percent.
Pause on that comparison, because it is the whole article. Same chocolates, same restaurant, same server, same check. Delivered as policy, two pieces were worth about 2.7 points of the bill over nothing. Delivered as a gesture - one piece, a pause, a turn back, a second piece - the same candy was worth about 4 points. The turn-back beat the policy by a margin the study's own contrasts call significant, and nothing about the candy changed. What changed is that it arrived looking like a decision somebody had made about you.
The study is not folklore. David Strohmetz of Monmouth University, with Bruce Rind, Reed Fisher and Michael Lynn, published it in the Journal of Applied Social Psychology in 2002, under the title "Sweetening the Till: The Use of Candy to Increase Restaurant Tipping." The first experiment was simpler: parties that found one chocolate riding along with the check tipped an average of 17.8 percent of the bill, against 15.1 percent for parties that got none. Real restaurant, real diners, eighty tables in the second experiment alone.
The obligation forms before the ask
The candy was never doing the work. Pull the results apart and the active ingredients are three. The give has to arrive before the ask - the check is sitting on the table, unpaid, when the chocolate lands. It has to be unrequested. And it has to read as chosen for you, not issued to tables generally. Hit all three and people repay; the candy experiment isolates the third one alone, and it is worth real money.
The cleanest demonstration I know is older. In 1971, Dennis Regan ran an experiment at Cornell in which subjects rated paintings alongside a supposed second subject, who was really Regan's assistant - every retelling of the study calls him Joe. During a break, Joe left the room and came back with two bottles of Coca-Cola, setting one down with, "I bought one for you, too." Half an hour later, Joe mentioned he was selling raffle tickets for a new car at twenty-five cents apiece, and would the subject buy some? "Any would help, the more the better." The subjects who had taken the Coke bought twice as many tickets as the ones who had not. The detail I find more interesting is the one Regan was not expecting: on questionnaires, the subjects who said they disliked Joe bought just as many of his tickets as the ones who liked him - as long as the Coke had come first. Obligation is not affection. It does not need you to be liked. It needs the sequence. Robert Cialdini, who retold the study in Influence, notes that the study ran in the late 1960s, when a Coke cost a dime - and the average subject who took one still bought two tickets at twenty-five cents apiece.
Cialdini's favorite nonprofit number tells the same story at mailbox scale. The Disabled American Veterans mailed a plain appeal for years and got about an 18 percent response. Mailings that included a small, unrequested gift inside the envelope - sheets of personalized address labels, worth cents - pulled about 35 percent. Nearly double, from a gift that was in the donor's hand before any decision about the ask had been made.
The sequence, run backwards
Now look at what marketing actually spends its money on. Thank-you pages after checkout. Coupons that arrive with the order confirmation. Loyalty points that start accruing after the first purchase. Welcome kits sent to people who just signed up. All real, all fine for the relationship - and all worthless as reciprocity, because the ask already succeeded. The debt you wanted to create is the debt you already collected. Gratitude after a transaction is receipt-reading. The obligation that moves strangers exists only in the window between give and ask, and only an open ask can collect it.
When marketing does give before the ask, it tends to buy the wrong variable: size. Fatter discounts, richer free tiers, bigger swag. The candy experiment measures exactly how little size matters - the second chocolate added nothing while it looked like policy, and the tables that got two pieces merely because the rules said so still tipped less than the tables that watched a person come back for them. What scales the effect is not the amount given but the timing and the person-ness: before, unrequested, for you. A white paper behind an email form fails all three at once. The form is the ask, so the "gift" comes after it. The email is the price, so the gift is barter. The same PDF downloads identically for everybody, so nobody is chosen. We call it lead generation. The diner calls it a menu.
What the board gives away
A street performer worked this out centuries before the journals did. The song comes first - the whole song, unasked - and the hat goes around after. On Bid Spotlight, that sequencing is the product, on purpose. The bid buys the ask: a tile near the top of a board that people scroll to find things they did not know existed. The give side is the part most paid-placement boards never budget for. Every listing a buyer places here ships with an auto-made promo video - their product, their site - before they have had to do anything else. This blog is the same bet, ungated, because that is the address-label end of the operation.
If you are the one bidding, audit your own side of the sequence before you raise your number. The tile is your ask. Your give is whatever your site hands a stranger who owes you nothing, before any signup, any email, any price. Count it honestly; most sites hand over a headline and a form. A bid lands harder on a table that already got its chocolate.
Count what you give away before you ask for anything, then claim your spot on the board.