## PROBLEM READ The immediate problem is not “low traffic.” It is a conversion decline: roughly similar visitor volume is producing fewer purchases. Two meaningful store changes happened shortly before the decline: * product photos changed; * free-shipping eligibility moved from $30 to $45. The goal is therefore to identify which change, if either, is suppressing conversion before spending more to acquire traffic that currently converts poorly. ## EVIDENCE STATE Reported evidence: * Sales have been down for three weeks. * Traffic appears roughly normal. * Fewer visitors appear to purchase. * Product photos were changed about a month ago. * Free-shipping eligibility increased from $30 to $45 around the same period. * More Instagram posting did not resolve the decline. * A 10% weekend discount temporarily improved sales; performance declined again afterward. * Testing budget is capped at $100 and the shop is operated by one person. Reasonable inference: * The problem is probably downstream of traffic acquisition. * The temporary discount response suggests purchase economics or perceived value may matter, but it does not prove the shipping threshold caused the decline. * Because the photo and shipping changes overlap in timing, neither can currently be isolated as the cause. Decision-critical unknown: * Does removing the higher shipping threshold recover conversion while everything else stays approximately constant? A secondary unknown is whether the new photos reduced product-page effectiveness. That does not need to be tested simultaneously. ## STRATEGY ROUTE EXPERIMENT. There are at least two plausible explanations, both are inexpensive to reverse, and changing several things together would make the result harder to interpret. A small controlled rollback has better information value than more advertising or another discount campaign. ## BEST NEXT MOVE For one normal sales cycle, temporarily return free-shipping eligibility from $45 to $30 while leaving the current product photos unchanged. Do not simultaneously: * change photos, * run another discount, * substantially increase advertising, * redesign product pages. This is the strongest first test because it changes one consequential variable, costs little to run, is easy to reverse, and directly tests whether additional checkout economics introduced around the decline are suppressing purchases. Before changing it, record four numbers from the recent three-week period if your shop platform exposes them: 1. visitor/session count, 2. orders, 3. conversion rate = orders ÷ visitors, 4. average order value. Also record the same metrics from a comparable period before the two store changes if available. No paid experiment budget is required unless the lower shipping threshold generates additional shipping expense. ## ACTION LADDER 1. Now Run the $30 free-shipping threshold for about 7 days or long enough to cover your normal weekday/weekend pattern. Keep the rest of the store materially unchanged. 2. If the expected signal appears If conversion clearly rebounds, keep the lower threshold temporarily and evaluate its economics. Compare additional orders and revenue against the extra shipping cost. You can then test whether an intermediate threshold such as $35 or $40 preserves most of the conversion recovery while improving margin. 3. If it does not Restore the $45 threshold. Do not keep repeating discounts. Next, test the other major change: restore the previous product photos for a representative group of important products while leaving shipping unchanged. Compare product-view → add-to-cart and purchase behavior if those metrics are available. ## CHECKPOINT Use the pre-change conversion rate as the reference point rather than raw sales dollars alone. A useful success rule: > The shipping rollback should recover a substantial portion of the conversion decline without creating enough additional shipping expense to erase the resulting gross-profit improvement. For a simple working threshold, treat recovery of roughly half or more of the lost conversion gap as a strong enough signal to continue investigating shipping policy. Example: * Before changes: 3.0% conversion * Recent period: 2.0% * $30 shipping test: 2.6% The lost gap was 1 percentage point; recovering 0.6 points would be meaningful evidence that the shipping change matters. A weak result would be something like 2.0% → 2.1%, especially if normal daily variation could explain it. Do not require statistical sophistication for a small shop, but avoid judging the experiment from one unusually good or bad day. ## SWITCH RULES * IF conversion materially recovers after returning to $30 → investigate the best profitable shipping threshold rather than advertising harder. * IF conversion does not materially improve → restore $45 and test the photo change separately. * IF both tests fail to explain the decline → broaden the investigation to checkout friction, pricing, product mix, stock availability, traffic quality, device/browser problems, and other store changes during the same period. * IF conversion improves only during discounts → investigate price/value perception rather than continuing permanent promotions. * IF lowering the threshold increases orders but reduces total contribution profit → do not treat conversion improvement alone as success. * IF new evidence contradicts the working explanation → stop protecting the original hypothesis and reframe the problem. ## WATCH OUT The biggest mistake would be changing photos, shipping, discounts, social posting, and advertising simultaneously. Sales might improve, but you would still not know why. The 10% discount is useful evidence, but not proof that your regular prices are wrong. It altered the customer's total economic proposition, just as the shipping threshold did. I also have not inspected your store analytics or historical data, so the diagnosis above is a testable working strategy rather than an externally verified root cause.