Which sign, link or post brings you the best members?

Joining costs a customer an email address and a 6-digit code, and adding the card is one more tap. None of it takes a purchase. Everyone who walks past sees a sign on the door, including people who are curious for a minute and never come in again. Only people who are paying see a sign at the counter. The first counts footfall and the second counts customers.

That is reasoning, not a finding: we found no study that measures signs, links or posts against each other. The nearest sound evidence is about a different route. Researchers tracked about 10,000 customers of a German bank for almost three years: 5,181 who had arrived through its referral program in 2006 and a sample of 4,633 who arrived by other routes that year. The referred customers had "a higher retention rate, and this difference persists over time", and their average value was "at least 16% higher than that of a nonreferred customer with similar demographics and time of acquisition". This is one bank, and a referral is not a sign. It shows that people who look alike on the day they join can go on behaving differently depending on how they came. Only your own figures can say which of your routes does that.

A member has one source, fixed at the moment of joining. If two placements share a code, nothing can separate them afterwards.

Signs. Once your program is live, QR Materials opens with Ready-made QR codes: one code for each place a code is used (General, Counter, Standee, Window, Receipt and Online), each already live and opening the same join page. "Scans and joins are counted for each one, so Performance shows which place brings people in." Your own adds a category for a place the list lacks, and its code "is live at once". A new place needs its own code. Materials that share a code share one line, and the Performance view says of them: "Individual printed copies cannot be distinguished."

The category's name is how you tell the lines apart, so name a category of your own for exactly where it will stand: "Door, Jalan Telawi" tells you more than "Poster". With several outlets, add one for each counter. To print a designed sign with a ready-made code, choose One I already have on the sign's QR code page; the page calls it "Right for a reprint or a new look for the same place: scans are counted together." Its own new QR code is the other choice. It also offers Outlet ("Choose one to see this outlet’s joins separately."), but a new code is checked by native.cards before it goes live, where a ready-made one is live already.

Performance, the second view in QR Materials, counts Qualified views and memberships for each QR code and warns: "Period counts describe separate observations, not a matched conversion funnel." It tells you that a sign is being scanned. Whether those people came back is on the Campaigns page.

Links. Under the table in Campaigns is Create a tracked join link. Choose Where you will share it (Instagram, Facebook, TikTok, Google, WhatsApp, Email or Print) and what it is (Paid ad, Post or story, Message, Newsletter or Poster or flyer). Under Name this campaign, give the push a name such as raya-2026, then copy the link. The builder only labels the link. You still post or send it yourself. Members who join from it, the page says, "get their own line under Where members came from, with the visits and spend that follow", shown as raya-2026 · Instagram, post or story. Give each post or push its own name, because links that carry the same tags share a line.

Anyone who joins through an untagged link lands in a general join line and cannot be sorted afterwards. Members who came by a friend's invitation appear as Member invites (referrals). Lines for Instagram and Facebook Auto-Replies and for ads appear only where those connections are switched on for a business. Where they are, each post that sent a reply is listed with its likes beside its joins and spend. native.cards' own line on this is "Your most-liked post isn’t always your best seller".

Joined is "Members who joined your card program through this source in the chosen period." Came back is "Members from this source with two or more recorded visits in their first 30 days, and their share of joins." Read the count first, for how many returning members a source brought. Read the share second, for how many joins it took to get them. Visits, Spend and Rewards used cover the same first 30 days. Spend has nothing to add up on a card that counts only stamps or visits, because no amount is recorded. Wallet Cards is "Members from this source who added their card to Apple Wallet or Google Wallet." The period can be 30 days, 90 days or 12 months.

Every member is given the same 30 days, so a sign that has been up for a year is not flattered against one put up last month. The price is the label Still counting. It marks a line whose newest member joined less than 30 days ago, so its numbers can still rise. A sign that is still bringing people in wears the label all the time, because its newest member is always recent. Read its figures as "at least this many". For a one-off push, wait until the label goes.

At the top, Best return rate names "The source whose members came back most often, among sources with at least 5 joins". Five is the page's floor for naming a leader. It is not enough to act on.

The counter sign starts one visit ahead. Came back needs two recorded visits. A member who joins while paying, and whose purchase staff then record, has the first within a minute. A member who joins from the door or a link starts at nought. In the example, if staff recorded every joining purchase, 25 of the counter's 70 visits are those purchases. Take them off and 45 later visits remain, against 30 for the door. The counter still leads, from fewer than half the members, but by less than the Came back column suggests. It also reaches people who were customers already, which is part of why they return: the sign found them, it did not make them. Where you can, compare counter signs with counter signs and links with links.

It records where someone joined, not where they heard of you. A customer who sees your post on Tuesday, walks in on Saturday and joins at the counter belongs to the counter sign. The post gets nothing, and Campaigns cannot tell you what made anyone walk in.

Change one thing. Two signs that differ in place, headline and design cannot tell you which difference mattered. Print the same design twice, each with a different place's code, which is what a test of two places needs.

Use the same weeks. A sign that was up over a festive fortnight is not comparable with one that went up afterwards, so put both out on the same day. If you move a sign, give it the code of its new place, or its line will mix two places.

Wait for enough joins. By the rough table in How to test whether your program makes a difference, two groups of about 30 can differ by 25 points through luck alone. The example's gap is 41 points, so it stands. A gap of ten would not.

These are readings, and any of them may be wrong. Change one thing and let the next month's lines say whether the reading was right. The remedies have articles of their own: How to get customers to join at the counter, How to get new customers to come back a second time and How soon should customers earn their first reward?

Do not take a sign down because it loses on Came back. The door sign in the example brought 30 visits from people who joined on their way past. Stop counting its joins as success, and keep an eye on its Wallet Cards figure. Members who joined without adding the card cannot be reached by a notification, because notifications travel through the card.