AI can work out colour par levels and reorder points from your usage data, but it cannot count the shelf for you. Give it eight weeks of tubes used per shade and your supplier's lead time, and it returns par levels, reorder points and a draft order sheet that you check against a physical count before sending.
The real question is whether you have usage data. Many salons hold sales data, meaning services booked, but not shade-level usage, and a service count cannot tell you whether the tube that went was a 6N or a 7.1. The fix takes about ten minutes a week: count the shelf. Opening count plus deliveries minus closing count is your usage. Keep the sums in a spreadsheet, because an assistant that reasons well can still slip on arithmetic.
Usage first: three ways to know what you really use
Everything downstream depends on the usage figure, so choose the method you will actually keep up.
| Method | Effort | Good for | Weak spot |
|---|---|---|---|
| Weekly shelf count: opening plus deliveries minus closing | About 10 minutes, same time every week | Real usage by shade, including waste | A skipped week breaks the chain; counts must be by shade, not by box |
| Formula cards multiplied by grams per service | Needs weights recorded for each formula | New shades, new stylists, sanity checks | Ignores leftover mix, so it runs below real usage |
| Software that deducts stock as a service is recorded | Set-up once, then automatic | Salons already using the software for every service | Only as good as stylists' discipline in recording |
The weekly count sheet itself is small: for each shade, the opening count, tubes delivered, the closing count and the usage that falls out. Three illustrative rows:
| Shade | Opening | Delivered | Closing | Used |
|---|---|---|---|---|
| 7N | 5 | 8 | 7 | 6 (5 + 8 - 7) |
| 6.1 | 3 | 0 | 1 | 2 |
| 9N | 4 | 0 | 3 | 1 |
Count tubes, not boxes. Count a part-used tube as a whole one when about half or more is left, and as nothing when less is left. The exact rule matters less than keeping to it every week.
I would start with the weekly count, because it needs nothing but a clipboard and it captures everything that leaves the shelf, including the mix that goes down the sink. If you want a faster way to do the count itself, the stocktake tutorial with a phone scanner covers it. The software route is real but narrower than it sounds. Phorest's page lists stock deduction (recording what was used in a treatment, including partial amounts), minimum and maximum stock levels and supplier order forms, and it tracks professional stock separately from retail stock. Fresha's help centre lists a low-stock level and a reorder quantity you set per product, with a low-stock notifications switch, and its 2022 announcement of the feature adds a Monday-morning summary and auto-filled purchase orders, so confirm those details as they work today. None of these pages mentions tracking colour by shade, so ask for a demo with your own shade list before you rely on either.
Three numbers per shade: usage, reorder point and par level
For each shade you need three numbers, and confusion between the last two causes many ordering mistakes.
- Weekly usage: tubes used over the last eight weeks, divided by eight.
- Reorder point: the count at which you must order now, even mid-week. It equals the usage you expect during the supplier's lead time, plus a safety cushion.
- Par level: the count you order up to at each weekly review. It equals the usage you expect from now until the next delivery could arrive (review period plus lead time), plus the same cushion.
My starting rule for the cushion is half a week of usage, rounded up, with a minimum of one tube. Anything is better than no cushion, and you can tune it later: a shade whose absence loses a client gets more, a slow shade you can get next day gets less. The order quantity at each review is the par level minus what you hold (on the shelf plus already on order), never below zero. For the general method behind these terms, see how to use AI to set reorder points and prevent stockouts. The rest of this tutorial is the salon version.
Ten colour lines worked through from eight weeks of counts
The main worked example is an illustrative five-chair salon. It counts the colour shelf every Monday morning before any mixing, orders on Monday for delivery two days later, and pays $9 a tube. Eight weeks of counts gave the usage figures below, and today's Monday count is in the "On hand today" column.
| Shade | Used in 8 weeks | Weekly usage | Safety | Reorder point | Par | On hand today | Order |
|---|---|---|---|---|---|---|---|
| 4N | 14 | 1.75 | 1 | 2 | 4 | 3 | 1 |
| 5N | 38 | 4.75 | 3 | 5 | 10 | 6 | 4 |
| 6N | 46 | 5.75 | 3 | 5 | 11 | 4 | 7 |
| 7N | 52 | 6.5 | 4 | 6 | 13 | 5 | 8 |
| 8N | 30 | 3.75 | 2 | 4 | 7 | 5 | 2 |
| 9N | 12 | 1.5 | 1 | 2 | 3 | 4 | 0 |
| 6.1 (ash) | 18 | 2.25 | 2 | 3 | 5 | 2 | 3 |
| 7.1 (ash) | 24 | 3.0 | 2 | 3 | 6 | 6 | 0 |
| 8.3 (gold) | 10 | 1.25 | 1 | 2 | 3 | 1 | 2 |
| 10.1 (very light ash) | 6 | 0.75 | 1 | 2 | 2 | 5 | 0 |
| Total | 250 | 31.25 | 20 | 34 | 64 | 41 | 27 |
Take 5N to see where each figure comes from. It used 38 tubes in eight weeks, so 4.75 a week. Safety is half of that, 2.4, rounded up to 3. The lead time is two days, so expected usage while waiting for a delivery is 4.75 times 2 over 7, which is 1.36, rounded up to 2. Reorder point is 2 plus 3, so 5. Par covers the review period and the lead time, nine days in all: 4.75 times 9 over 7 is 6.1, rounded up to 7, plus 3, so 10. With 6 on the shelf, the order is 4.
Now read the table the way you would on Monday morning. Four shades are below their reorder points: 6N (4 against 5), 7N (5 against 6), 6.1 (2 against 3) and 8.3 (1 against 2). Without reorder points, those would have run low during the week and been spotted when a formula could not be mixed. At 0.7 to 0.9 weeks of cover each, they have the thinnest cushion going into the delivery, and one busy day could empty a shade. The order totals 27 tubes, or $243. Meanwhile 10.1 holds five tubes against a weekly usage of 0.75: that is 6.7 weeks of cover, so nothing is ordered and the five tubes are $45 sitting on a shelf. 9N has 2.7 weeks of cover, a lesser version of the same story.
The stock value tells the same story. Holding every shade at par costs 64 tubes, or $576. The shelf today holds 41 tubes, $369, yet four shades are short. A shelf can be underfilled and overfilled at once, and par levels fix both by putting the stock where the usage is.
One more thing an eight-week average hides is a trend. Here are 7N's weekly counts in this illustration: 5, 5, 6, 6, 7, 7, 8, 8 tubes, which total 52 and average 6.5. The first four weeks averaged 5.5 and the last four 7.5, a rise of about 36%. Ask the assistant to compare the two halves of the window for every shade and flag any that moved by more than 15%. Where the rise looks real, for instance a stylist building a blonde clientele, base the sums on the last four weeks instead of all eight. Recalculated at 7.5 a week, the par for 7N becomes 14 and the reorder point 7, one tube more at each, which is the size of cushion that stops a quiet drift turning into a missed formula.
The spreadsheet formulas, so the maths never depends on a chat
Have the assistant write the formulas, then keep the sheet. The same logic in a spreadsheet takes five formulas per row. Put the lead time (2) in cell L1 and the review period (7) in cell L2, shade names in column A, eight-week usage in column B, today's count in G and anything already on order in H.
C2 =B2/8 weekly usage
D2 =MAX(1,ROUNDUP(C2*0.5,0)) safety tubes
E2 =ROUNDUP(C2*$L$1/7,0)+D2 reorder point
F2 =ROUNDUP(C2*($L$1+$L$2)/7,0)+D2 par level
I2 =MAX(0,F2-(G2+H2)) order quantityCopy the row down for every shade. Change L1 when your supplier's lead time changes and every row updates. The assistant is useful for writing this and for explaining it, but the formulas are the engine: they give the same answer every Monday, and a chat does not promise to. The tutorial on why AI stock forecasts go wrong lists eight mistakes small businesses make with AI forecasts.
A prompt that builds the sheet, and what to check in the answer
If you would rather paste a table into a chat and get a table back, state every rule, because an unstated rule gets a plausible default.
I run a hair salon. Below is a table of colour shades with tubes used in the last
8 weeks and today's shelf count. Settings: supplier lead time 2 days, I review
and order every 7 days, tubes cost $9 each.
Rules:
1. Weekly usage = tubes used / 8.
2. Safety stock = half a week of usage, rounded up, minimum 1 tube.
3. Reorder point = usage over the lead time (rounded up) + safety stock.
4. Par level = usage over lead time + review period (rounded up) + safety stock.
5. Order quantity = par level - (on hand + on order), never below zero.
6. Flag every shade whose on-hand count is below its reorder point.
7. Add weeks of cover (on hand / weekly usage) and flag any above 4 weeks.
Show the table, then show your arithmetic for the three most-used shades.
[paste table]Check three things in what comes back. First, rounding. A typical slip is rounding 7N's 8.36 down to 8 before adding the safety stock, which gives a par of 12 instead of 13. On any shade with a fractional result, rounding down leaves the shelf a tube short. Second, the review period. Some answers calculate par over seven days instead of nine, so every par comes out low by about two days of usage. Third, the flags: count the shades below reorder point yourself, which here is four, and confirm the assistant found the same four. If it did not, the table is wrong somewhere, so do not use it until you find where.
From order quantities to a supplier order sheet
Once the quantities are right, the assistant can format them into the layout your supplier expects. Paste the last invoice so it can copy the product codes, and ask it to group by supplier. An illustrative result, with invented codes:
| Supplier code | Shade | Quantity | Unit cost | Line total |
|---|---|---|---|---|
| PC-04N | 4N | 1 | $9.00 | $9.00 |
| PC-05N | 5N | 4 | $9.00 | $36.00 |
| PC-06N | 6N | 7 | $9.00 | $63.00 |
| PC-07N | 7N | 8 | $9.00 | $72.00 |
| PC-08N | 8N | 2 | $9.00 | $18.00 |
| PC-0601 | 6.1 | 3 | $9.00 | $27.00 |
| PC-0803 | 8.3 | 2 | $9.00 | $18.00 |
| Total | 27 | $243.00 | ||
Four checks before you press send. Does the quantity column add to the 27 tubes your sheet says? Does each code match the last invoice for that exact shade? Similar shade numbers are where this goes wrong: 7.1, 7.01 and 7.11 are three different shades in many ranges, and an assistant matching a name to a code can pick the wrong sibling. Does the total clear your supplier's minimum order or free-delivery threshold, and if it falls short, which fast-moving shade would you top up to reach it? And is the order inside the supplier's cut-off for the delivery day your lead-time figure assumes?
Grams per service: a cross-check that also covers new shades
Counts tell you what left the shelf. Formula cards tell you what the services should have used, and the gap is a useful signal. Tube sizes vary by brand, so read the size printed on yours. To turn grams into tubes, weigh one full tube and one empty tube on the salon scale and take the difference; this illustration assumes 60 ml tubes holding about 60 g of colour. Mixing ratios vary too: one manufacturer's own formula page mixes its permanent colour 1:1 with a 6% (20 vol) developer and a clay lightener 1:2 with a 12% (40 vol) developer, and other lines use other ratios, so read your own chart. Then weigh your formulas for a fortnight and average them. An illustrative week:
| Service | Per week | Colour per service | Colour per week |
|---|---|---|---|
| Root retouch | 26 | 38 g | 988 g |
| All-over colour | 7 | 85 g | 595 g |
| Gloss or toner | 8 | 25 g | 200 g |
| Total | 41 | 1,783 g |
1,783 g is 29.7 tubes at 60 g, against 31.25 tubes counted. The counted figure is 5% higher, which is the sort of gap leftover mix produces. My working rule is to start asking questions at around 15 to 20%: it points to unrecorded services, a lot of bowl waste or a miscount. The same table prices your colour. At $9 for 60 g the colour costs $0.15 a gram, so a 38 g root retouch carries $5.70 of colour before developer, which is the figure to check against your service price (see writing a salon service menu and price list with AI).
New shades are where this table helps most. With no counts yet, estimate how many services will use the shade, multiply by the grams, and review at week four. A simpler fallback is to borrow the weekly usage of the nearest existing shade, halve it, and review at week four.
Developer, lightener and the lines that behave differently
Developer is bought in litres and used alongside nearly every colour service, so give each strength its own row, for example 20 vol, which is 6%. At a 1:1 ratio with 60 ml tubes, 31 tubes of colour a week implies about 1.9 litres of developer for colour services alone, before any lightening or glosses at other ratios. Lightener is usually weighed in grams from a tub, so track it in tubs and keep its weekly usage separate from colour. Toners and glosses are many shades with small volumes each, so give them one shared sheet with a higher safety minimum rather than precision you cannot support. For shelf life, follow the manufacturer's guidance and the batch code on each pack, and put new deliveries behind the old stock so the oldest tube goes first.
Discontinued shades, box sizes and other cases that break the sums
- A shade is discontinued. Set its par to zero, use up what is on the shelf and ask the supplier for the replacement. The sheet will otherwise keep ordering it.
- Seasonal peaks. If you kept counts last year, ask the assistant for the ratio between usage in the three weeks before each peak and your yearly average, then raise par for those weeks by that ratio and order a week early. If you have no history, do not guess a figure; raise safety stock on your top six shades instead.
- Pack sizes. If a supplier sells boxes, round each order up to a whole box and say so in the prompt. Otherwise the order sheet will ask for three tubes of something sold in packs of six.
- A stylist changes line. Usage moves from one range to another. Zero out the old line's par as the stock drains and rebuild the new range from two or three weeks of counts.
- A shade is out of stock at the supplier. Log it. Two missed deliveries in a quarter is a reason to hold more of that shade, or to change supplier. The tutorial on finding dead stock and slow movers covers the opposite problem.
The Monday routine, start to finish, in about 25 minutes
A routine only works if it fits into the week without effort. This is the sequence for the salon above:
- Count the shelf before any mixing (10 minutes). Fill in the closing count for every shade, including the developer and lightener rows.
- Enter last week's deliveries and work out usage (2 minutes). Opening plus delivered minus closing, by shade.
- Roll the window forward (2 minutes). Drop the oldest week, add the new one, so the sheet always holds the latest eight weeks.
- Read the flags (3 minutes). Shades below their reorder point, shades above four weeks of cover and any shade whose usage moved by more than 15%.
- Draft and check the order (5 minutes). Let the assistant format it, then check the quantity total, the product codes and the minimum order value.
- Send it before the supplier's cut-off (3 minutes). On delivery day, tick each line against the box and write the delivered counts into the sheet for next Monday.
That is about 25 minutes a week for a salon of this size, and most of it is the count. If a week is missed, do not skip it: count anyway and enter two weeks of usage divided by two, so the window never has a hole in it.
How to tell after six weeks that it worked
Write down four numbers for the six weeks before you start, and the six weeks after. Missed formulas: the times a stylist could not mix a formula because a shade was out; a sticky note on the shelf is enough. Emergency orders: mid-week orders placed because something ran out. Weeks of cover: on hand divided by weekly usage, averaged over your top ten shades. Stock value: shelf count times unit cost, counted on the same weekday.
The targets I would set are zero missed formulas on the six busiest shades, average cover between one and two weeks, no shade above four weeks, and one or fewer emergency orders in the six weeks. If cover creeps above two weeks, your safety stock is too generous. If missed formulas continue on one shade, give it a larger cushion rather than changing the whole system. The money side is modest but real: the worked example holds $576 of stock at par, and every tube you stop buying for a shade like 10.1, with weeks of cover to spare, is cash back in the bank. The hair salon business plan tutorial covers chairs, prices and cash flow.
That is the whole routine: a ten-minute count, a spreadsheet that does the same sums each week, an assistant that drafts the order sheet and flags what looks wrong, and a person who confirms the order against the shelf. The assistant earns its place by saving the typing and spotting the odd line. It does not replace the count.
Questions salon owners ask about colour stock and AI
How many weeks of usage data do I need?
Eight weeks is a workable start because it smooths out a quiet fortnight and a busy one. Use twelve or more if your trade is seasonal, and refresh the sheet monthly. For a new salon with no history, estimate from service counts and formula cards, hold generous safety stock, and replace the estimates with real counts as they arrive.
Can my booking software track colour usage by shade?
Some software tracks stock levels and low-stock alerts, and Phorest's page describes recording what was used during a treatment, including partial amounts. Neither that page nor Fresha's inventory help pages mention tracking colour by shade, so ask the vendor to demonstrate it with your own shade list before you pay.
Should AI place the order for me?
No. Let it draft the order sheet, then have a person confirm it against a physical count of the shelf, check the product codes against the last invoice and send it. A wrong code or a miscounted shelf costs more than the few minutes the automatic version saves.
How much safety stock should I hold?
My starting rule is half a week of usage, rounded up, with a minimum of one tube, so a shade you use six tubes a week keeps three in reserve. Raise it for shades where a missed formula loses a client or where the supplier's deliveries are unreliable. Lower it for slow shades you can get within a day.
Further reads
- 5 AI Tools for Salons: A Stack for Solo Stylists and Small Teams — A wider stack of AI tools for solo stylists and small teams.
- Fresha Pricing: Subscription, Marketplace Fees and Add-Ons Explained — What Fresha charges, in case its stock tools matter to you.
- Vagaro Pricing for a Small Salon: Calendars, Add-Ons and AI Features — Vagaro's plans and add-ons for a small salon.
- AI Assistant for Hair Salons: ChatGPT, Claude, Gemini or Copilot? — Choosing between ChatGPT, Claude, Gemini and Copilot in a salon.
- Boutique Inventory Management: AI for Size Runs and Reorders — Inventory and reorders for stock that comes in sizes.
- How Small Shops Can Use AI to Forecast Stock and Reorder — Forecasting stock and reordering in a small shop.
- Hardware Store POS Systems: Which Features Matter and Where AI Fits — A feature checklist for an independent hardware store POS: SKU depth, trade accounts, special orders and reorders, plus what AI really does in each.
- How Much to Charge for Hair Extensions: Fitting, Hair and Aftercare — A worked price table for tape-in, beaded and keratin extensions: hair cost, fitting time, maintenance visits and aftercare, modelled with AI.
- Salon Pricing Calculator: Build One With AI From Your Real Costs — Cost per chair minute, product cost per colour and a margin target, built into a ChatGPT-made spreadsheet you can check line by line.
- Salon Retail Sales: Write Aftercare Product Recommendations With AI — Product cards, a service-to-product map, a safe prompt and a claims check, so aftercare texts sell honestly without health claims.
- Tyre Stock Excel Sheet: Track Sizes and Reorder Points With AI — A tested tyre stock sheet: size-and-brand matrix, reorder-point formulas, seasonal demand, slow-mover and DOT-age flags, plus three jobs for AI.
- Salon Patch Test Form Checklist: Reminders and Records With AI — Form fields, reminder texts timed to the 48-hour check and a record-keeping routine for patch tests, with AI drafting the wording and people judging reactions.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Phorest's stock and inventory feature page; Fresha's inventory help centre and its 2022 low-stock notification announcement; Schwarzkopf Professional's colour-melt formula page (1:1 with a 6% / 20 vol developer). Stock figures in the worked example are illustrative. Checked October 2026.