Scale · capability

Built for millions of SKU-locations.

The same connected plan, from a single brand to a multi-fascia group — reconciled live at the scale your business actually runs.

TightlyPlanning·Forecast engine · liveLast sync · just now

Plan every cell. Refresh inside the window.

SKU × location × day across the whole catalogue — recomputed while you're still trading, not overnight.

Forecast engine · recomputed 14 min ago
4.2MSKU × location × day
12,480SKUs
38Locations
<2sRecompute
11.4%Err
Full-resolution, always fresh
12.5k
SKUs live
38
locations
daily
grain
<2s
recompute
Recomputed just now
4.2M cells ›
Yarrow · London · dailybest-fit14m ago
Iris (new SKU) · 3 locscatalogue lift14m ago
The numbers behind it

Why scale decides if you plan at SKU or category.

14 pts

Full-price sell-through gap between retail leaders and the industry — most of it down to how the group plans together at scale.

Source: Directional industry benchmark

~53%

Share of unplanned markdown cost attributed to upstream decisions in the group's biggest brands.

Source: Coresight industry research

$13.2M

Total exposure on the group risk register in a typical audit cycle — visible live in CFO reports.

Source: Representative multi-brand retailer

The problem today

What breaks before scale runs to plan.

TightlyPlanning·Scale · Plan resolution
Plan resolution
Cat × Month
detail aggregated away
Refresh
weekly batch
recomputed overnight
Sparse / new SKUs
dropped
too little history
Cells a sheet holds
~3.2K
before it corrupts
Plan freshness
stale by Mon
batch finishes late
Plan coverage · what a spreadsheet can hold
2 of 5 tiers out of reach
Category × Month~180 cellsSubcategory × Month~1.4K cellsSKU × Month~3.2K cellsSKU × Location~120K cellsSKU × Location × Day4.2M cellsas far as spreadsheets reach ↑the grain you actually trade at ↓out of reach in a sheetcorrupts / slows past a fewthousand cells
Where the plan breaks · four failure modes
Planned at category × month
the SKU × location × day you trade at is aggregated away
Stale before the batch finishes
recomputed overnight, wrong by the time it lands
Sparse & new SKUs guessed
too little history, so they fall out of the plan
One method for everything
a single curve stretched across the whole catalogue
The refresh trap
Stale before it’s finished
By the time the weekly batch finishes, the plan is already wrong.
Diagnosis — this is a cardinality + refresh problem, not a spreadsheet-formula problemIllustrative · representative mid-market catalogue
How Tightly does it

Three steps from a million cells to a recompute in the trading window.

TightlyTightly · Forecast engine live· three stages from data to decisionLive
01Fit
CATALOGUE · ROUTED TO A METHOD FAMILYcatalogSTABLESEASONALTRENDSPARSE

A best-fit method for every series

Every SKU × location series is routed to the method that fits its shape — stable, seasonal, trending or sparse — so no single curve is stretched across the whole catalog.

best-fit per series
02Lift
NEW SKU · FORECAST LIFTED FROM LOOK-ALIKESlook-alike products3 weeks of historyforecast, lifted

Lift the long tail with a catalog-wide model

Sparse and brand-new SKUs have too little history to forecast alone. A catalog-wide model borrows the pattern of look-alike products to draw a real forecast line in from the population.

catalog-wide lift
03Recompute
RECOMPUTE · INSIDE THE TRADING WINDOW<2srecomputeHOURLY09:0010:0011:0012:00CELLS · FRESH4.2M cells · recomputed 14m ago

Recompute inside the trading window

The whole plan — millions of SKU × location × day cells — is recomputed as fresh signal lands, hourly, while you're still trading. No overnight batch to wait on.

hourly · <2s
See it run

Enterprise scale, without the lag.

The same connected plan, from a single brand to a multi-fascia group — reconciled live at the scale your business actually runs.

TightlyPlanning·Scale · Forecast engineRecomputed 14 min ago
Forecast cells
4.2M
SKU × location × day
SKUs
12,480
whole catalog
Locations
38
stores + DC + channels
Recompute
<2s
at full catalog
Freshness
14 min
hourly recompute
Best-fit method · per series
every SKU fitted automatically
Stable34%
Seasonal26%
Trend18%
Sparse / intermittent14%
New · catalog-wide lift8%
The sparse & new tail is lifted by a model trained across your whole catalog — the SKUs a single method can’t forecast alone.
Recompute log
live
Wide-Leg Pants · London
312k cells
14 min ago
Dresses · all locations
486k cells
14 min ago
Knitwear · Manchester
204k cells
14 min ago
New-season drop · 3 locs
58k cells
14 min ago
Every cell recomputed on Tightly’s engine — hourly, inside the trading window4.2M cells · <2s
Your agents

Meet your Supply agent

Works across every SKU-location at once — reconciling cover, lead time and open-to-buy in real time.

Meet the agents
Tightly agent
just now · within your limits
Live

Re-forecast ready — 3 categories have drifted from plan this week. Want me to stage the moves for your review?

Drifted vs plan · this weekΔ wmape
Tailored Trousers+9%8%
Woolly Layers−12%11%
Activewear+5%9%
Rebalance 240u DC → SFRe-baseline OTB Q3Hold buy on OCN-072
Stage movesReview firstLogged · audit ready
Customer outcome
We used to order stock based on vibes and prayers. Now we order based on data, and the difference in how the business runs is something none of us would go back from.
Consumer electronics
Director of Supply Chain
An enterprise brand
−32%

Reduction in forecast error

What this replaces

The engine you run today, and the one Tightly delivers.

Spreadsheet / legacy tool · plan grid
PLAN GRID · CATEGORY × MONTH~3.2k cells
CategoryMonthFcstWide-Leg PantsJul18.4kKnitwearJul9.1kDressesJul6.2k
category × monthovernight batchsparse SKUs droppedstale by Monday
A spreadsheet buckles at millions of SKU-location-day cells, so the plan rolls up to category × month and refreshes overnight. The grain replenishment needs is lost, and the numbers are stale before the batch finishes.
Tightly · Forecast engine · live
SKU × LOCATION × DAY · WHOLE CATALOGUErecompute <2s
SKULocDay fcstYarrow · Black · MLondon7/dBramble · Cream · SMcr4/dIris (new) · BlackDC3/d
SKU × location × dayrecompute inside the windowbest-fit per seriescatalogue-wide tail lift
Tightly’s forecast engine holds the whole catalogue at SKU × location × day — millions of cells — and recomputes inside the trading window, with a best-fit method per series and a catalogue-wide model lifting the tail.
FAQ

Questions buyers ask, answered straight.

Something not covered here? Talk to the team.

What scale has this been tested at?

Single brand: ~50k SKU-locations. Multi-fascia group: 2-5M SKU-locations across 4-12 brands. Live reconciliation works at both ends.

Can different brands have different policies?

Yes. Each brand has its own margin floors, OTB envelopes, cover targets, sign-off thresholds. The group only sees the roll-up; brand teams own the policy.

Does this satisfy audit / SOX requirements?

Audit log on every plan move, every override, every approval — exportable to standard formats. Most teams use it for SOX and internal audit on inventory planning decisions.

How does multi-currency work?

Plans run in the brand's reporting currency; group roll-up converts to the group's. Conversion rates are configurable — most teams set monthly at the cycle gate.

Enterprise scale. Single brand to multi-fascia group, one plan.

There's nothing to rip out. Tightly runs on your existing ERP, EDI, e-commerce and POS. Give us 30 minutes and we'll show it on your own categories.