Referral tracking in a CRM: what small teams actually need

Why referral tracking breaks when it lives in a spreadsheet

A referral motion has three moving parts: a customer mentions you to someone, that someone becomes a lead, and you need to know who sent them and whether it turned into revenue. Small teams usually track it in a shared sheet. Name of the referrer, name of the new contact, date. It works fine for about four months.

Then three things happen. The referral lives in the sheet while the deal lives in the CRM, so nobody can say whether any of it converted. The ask becomes manual, because remembering to email ten happy customers is a favour you do for two or three of them and then forget. And the thank-you stops, which matters more than it sounds, because the thank-you is the thing that produces the next referral.

The uncomfortable version: a referral you cannot attribute is a referral you cannot thank, cannot repeat, and cannot tell the referrer about. If a customer sent you three people and you never told them that, you have spent the single most valuable asset that customer gave you.

What a CRM can track about a referral

A CRM does two jobs here, and neither of them is the exciting one. It gives you attribution, and it lets you run the ask on a schedule instead of from memory.

• —A source field that survives. Put a fixed list on the contact record, referral as one value, plus the referrer's name as a second custom field. One record per person, with duplicate detection on import, so a contact referred by two different people does not become two rows and two half-credited referrals.
• —A form that captures the referrer at the moment it happens. Forms create the lead instantly, so a link or a code carrying the referrer's name can drop straight onto the record rather than onto somebody's clipboard.
• —The ask itself, as a campaign. Campaigns are segmented from the same records that hold the contacts and deals, with no list export and no sync, which is what makes it realistic to ask every quarter instead of never.
• —The thank-you, automated and timed. Automations run multi-step workflows with triggers, conditions and actions and no per-rule action meter, so the note that goes out three days after the deal closes happens whether or not anyone remembers.
• —One owner per referral. Ownership on every deal means the referred opportunity has a name against it, and the full activity history on that deal shows the referral source at the top of the record instead of in a separate spreadsheet.
• —A record of how it went. Reports on pipeline value, stage conversion, deal velocity and per-rep performance let you see whether referred opportunities behave differently from the rest, rather than assuming they are magic.

Notice what is absent: no reward engine, no unique link per advocate, no payout tracking, no commission ledger. A CRM gives you attribution and a repeatable ask. It does not give you the part where the referrer gets something.

What a referral tool does that no CRM does

Referral tools exist to run the reward and the ledger. ReferralCandy's own pricing page, read on 8 November 2026, lays out how that business is priced, and the shape of it tells you whether it fits your volume.

The mechanics matter more than the headline. Their success fee is a percentage of the purchase subtotal, after discounts and before taxes and shipping, charged only on net-new customers a referral brought in. Repeat orders after the first three are free of success fees, and there is never a fee on the reward you pay out to the advocate.

That structure is a volume business, and doing the arithmetic on their own published rates is the quickest way to know whether it is for you. At the Basic rate, a single referred sale of $1,000 carries a $105 success fee, which is more than the $39 monthly fee for the entire account. At the Grow rate, that same sale costs $35, and the base fee is $79. A business doing three referred sales a month is paying a referral tool's fixed fee to process $3,000 of referral revenue, and it would be cheaper to send three thank-you notes by hand.

The same arithmetic in reverse is why the tools look like a bargain at scale. At the Scale rate the fee is 1.5%, so a business doing 200 referred orders a month at $1,000 each is paying $249 fixed plus $3,000 in success fees on $200,000 of referred revenue. Against that, the CRM approach means somebody reconciles a referral ledger every month by hand. Which is cheaper depends entirely on your volume, and the pricing table is the most honest way to decide.

Honest caveats: when a small team should buy a referral tool

Five situations where buying the referral tool is the right call, and one where neither is.

• —Reward fulfilment is your bottleneck. If you are manually emailing vouchers, credits or free months, and you are losing track of who has been paid, that is bookkeeping you should not be doing by hand. This is the clearest case for buying the tool.
• —You need a unique link or code per advocate. If you have fifty people sending referrals, you need to know which link produced which order. A source field on the contact can tell you who referred them, but it cannot stop one advocate claiming another advocate's referral, and that is exactly what unique codes are for.
• —You are running affiliates as well as customers. ReferralCandy covers both, including commission strategy. BeigeCRM has no affiliate programme and does not pretend to.
• —Your referred volume already clears the fee. Do the arithmetic on the published rates above and compare the success fee against the admin time of reconciling by hand. Above a few hundred referred orders a month the tool is usually the cheaper answer.
• —You need fraud handling. Duplicate claims, self-referrals and fake orders are a category of problem that a spreadsheet does not solve at all.

And the case for neither: if you have not yet worked out who your happy customers are and why they are happy, a referral programme is a distribution decision that has not been made yet. Fix that first.

We also owe you the plain admission about our own product. BeigeCRM ships no referral tracking and no reward engine, and we checked our own features page on 8 November 2026 to be sure: no referral tracking, no per-advocate links, no payout ledger. What we give you is the substrate, which is custom fields, forms, campaigns and automations on top of one contact record and one deal record. That is enough to attribute a referral, run the ask and thank people on time. It is not enough to run a rewards programme, and if that is the part you need, buy a referral tool for it rather than trying to build one out of our features. And if the referral programme is really a partner or reseller network with territories, commissions and per-partner reporting, that is an enterprise shape with its own prices, so read our HubSpot comparison, Zoho comparison and Salesforce comparison before assuming a small CRM can carry it.

What to set up in the first fortnight

If the referral motion is real and the CRM is where you will run it, this is the order that works.

• —Pick one field, with a fixed list. Source, where referral is one value among several, and a second field for the referrer's name. Free text is how one brand ends up as Referral and another as referral.
• —Make the form carry the referrer. A hidden field or a code in the link drops the name onto the record before anyone can forget it.
• —Turn on duplicate detection before you import anything. A referred contact who is already in your system should update, not duplicate.
• —Build the ask as a campaign, not a favour. Segment from your own records, pick a date, and send it to people who have actually bought.
• —Automate the thank-you, with a delay, and name the referrer in it. That message is the mechanism that produces the next referral, and it should not depend on somebody noticing.
• —Decide what the field buys you. One number every month: how many referred contacts became opportunities, and what were they worth. If you cannot produce that, you are collecting attribution and not using it.
• —Record why referred deals are lost. Honest loss reasons on the deal tell you which referrers send buyers and which send browsers, which is the only part of this that compounds.

None of that needs a referral tool. It needs one field, one form, one campaign and one automation, and a decision about what happens when a referred customer actually buys something.