Run a Stocktake (choose the right scope)
A stocktake reports what is actually on the shelf; Alfred works out the difference and adjusts Shopify. The decision that matters most is the SCOPE, chosen by the card you start from — it is the only thing that decides which products the count is allowed to zero. Start narrow unless you genuinely mean to count everything.
Option 1: Specific Products — a spot check
RecommendedGo to Stocktakes & Adjustments and click 'Perform Stocktake: Specific Products'.
Choose the counting location and name the session (Alfred suggests one from the location, date and scope).
Add only the products you intend to count — Add from catalog, barcode scan, or a CSV of SKUs.
Enter the actual on-hand for each. This is the absolute count, not an increment. Typing 0 is a real count meaning 'the shelf is empty'.
Move to Review & Submit and check Count Variance and Variance Value for anything that looks wrong.
Submit. Nothing outside your list is touched, and nothing is written down for not being counted.
Option 2: Product Set — cycle count one vendor or category
Go to Stocktakes & Adjustments and click 'Perform Stocktake: Product Set'.
Choose the location, then narrow by vendor, product type, collection or tag.
Alfred builds the list and fills the count sheet with every matching product — you do not add them by hand.
Set Count visibility to 'Blind count' before counting starts if you don't want counters seeing the expected quantity.
Count. Anything you don't reach stays flagged as not counted rather than becoming a zero.
At Review & Submit, read the 'Not counted or input' card FIRST — that is the set at risk of being zeroed. Choose Zero or Ignore, and untick any row you want left alone.
If anything will be zeroed you must type a confirmation naming how many products it covers. Then submit.
Option 3: Whole Location — a wall-to-wall count
Go to Stocktakes & Adjustments and click 'Perform Stocktake: Whole Location'.
Choose the location. Every tracked product there is now in scope, and any of it can be written down.
Count through the location. Blind count is worth using here — this is the count that measures how accurate your records really are.
At Review & Submit, work through the 'Not counted or input' card carefully. On a whole-location count this can be a large set.
Type the confirmation and submit. Export the completed count as CSV for your records.
Pro Tips
- 💡Scope is the safety model. A Product Set count of one vendor cannot change another vendor's stock no matter what — which is exactly why it beats a whole-location count for routine cycle counting.
- 💡Adding a product to a count is not the same as counting it. An untouched row is treated as uncounted, never as a deliberate zero.
- 💡Watch Total net variance value, not just units. A count can be a handful of units off and thousands of dollars wrong.
- 💡Products with no cost recorded are excluded from the value figures and reported separately — if that number is large, your valuation is understated until you add costs.
- 💡Counting a product at a location where Shopify shows it untracked will activate it there and set the counted value.