# What do loyalty point farming bots look like in a store's logs? | Bot Traffic Index

Source: https://bottrafficindex.com/blog/2026-10-10-loyalty-points-farming-bots-logs.html

Oct 10, 2026 - Bot Traffic Index


Loyalty farming is the quietest form of bot fraud in ecommerce. No card testing, no login storms, no dramatic traffic spikes. Just a steady drip of new accounts earning signup points and redeeming them, at a scale that looks like a successful loyalty program until you look at who is actually buying.

## The signup pattern

Farming starts with account creation, and the log signature is volume without diversity. Hundreds or thousands of new accounts appear over days or weeks, each completing exactly the actions that earn points: account creation, email signup, maybe a first-purchase bonus trigger. The accounts never browse. They never add items to a cart and abandon them. They move through the point-earning checklist with machine efficiency and then go dormant until redemption.

The email addresses are the tell. They follow patterns: sequential, generated from the same domains, or drawn from disposable address services. The names are plausible but the behavior is not. Real new customers explore; farming accounts execute. In the logs, the difference shows up as session depth. A real signup session touches five to fifteen pages. A farming signup touches exactly the pages that pay points, in the same order, every time.

## The dormancy gap

Between earning and redemption, farming accounts go quiet. This is what separates them from ordinary bonus hunters, who are real people acting opportunistically. A real customer who signed up for the points might redeem them next week or forget about them entirely. Farming accounts redeem on a schedule, often in batches, because the operation behind them is working through an inventory of accounts systematically.

Watch for redemption clustering: dozens of accounts redeeming similar point amounts within a short window, shipping to a small set of addresses, or converting points to gift cards rather than products. Gift card redemption is the preferred exit because it is liquid and resalable. A loyalty program whose top redemption channel is gift cards, driven by accounts with no purchase history, is being farmed. The program is not failing; it is being harvested.

## Why it is hard to see in the dashboard

Loyalty farming looks like success in every standard metric. New account signups are up. Loyalty enrollment is up. Email list growth is up. The marketing team reports a thriving program while the finance team watches margin erode on redemptions that never produce real customers. The fraud is invisible at the aggregate level because every individual event, a signup, a redemption, looks legitimate. Only the pattern across accounts reveals it.

This is why the detection has to happen in the logs, not the dashboard. The signals are relational: shared IP ranges across "unrelated" accounts, identical user agent strings, signup-to-redemption timing that clusters too tightly, shipping addresses that repeat across accounts with different names. No single account looks fraudulent. The fraud is in the correlation, and correlation is what log analysis is for.

## What actually stops it

The effective defenses are boring and structural. Delay point availability so signup points cannot be redeemed instantly; farming economics depend on fast turnover. Require a real purchase before points become redeemable, which raises the cost of each farmed account above its expected return. Cap redemptions per household or shipping address, and enforce the cap. None of these stop a determined attacker completely, but they change the unit economics, and farming operations are businesses that respond to unit economics.

The deeper fix is to stop measuring the loyalty program by enrollment. A program with ten thousand farmed accounts and a program with one thousand real customers look identical in the signup chart and opposite in the revenue chart. Measure redemption by accounts with genuine purchase history, track the repeat purchase rate of point earners, and treat a spike in signups with no corresponding engagement as a fraud signal, not a marketing win. The program you want is smaller than the program you have. The logs will tell you which is which.

## Frequently asked questions

### What do loyalty point farming bots look like in logs?

High-frequency requests to account and rewards endpoints, many accounts created from the same networks, points accrued without corresponding purchases, and redemptions shipped to clustered addresses. Individually each looks plausible; in aggregate the pattern is mechanical.

### Why do bots farm loyalty points?

Points are currency. Farmed points get redeemed for goods, sold to others, or stacked with promotions for resale arbitrage. Loyalty programs are attractive targets because the rewards have real value and the fraud controls are often weaker than payment fraud controls.

### How can a store tell farming apart from enthusiastic customers?

Velocity and coordination. Real enthusiasts do not create dozens of accounts from one IP range or earn points at inhuman speed. Look for the mechanical signatures: identical timing, shared infrastructure, and accrual without purchase behavior.