Financial forecasting states where a period ends, and almost every method available for it works by extending a closed period forward. Sena supplies the one input a historical series omits: what consumers in each market are choosing this week and what would change it.
A financial forecast states three things: the revenue a period lands on, the cost against it, and the cash position that results. Three properties of the standard method leave the revenue line the weakest of the three.
The published techniques share one input. Percent of sales, moving average, regression, and straight-line extension each read a closed period and project it. The technique varies, and the evidence base holds constant, so a market turning this quarter reaches the forecast a quarter late.
Where the series is thin, a commercial view gets added on top. That judgement is frequently sound and rarely evidenced, so the forecast carries a number that survives review by seniority.
Forecast accuracy is usually reported as a single percentage. A regional figure of 95 percent conceals two markets at 80 and two at 110, and the plan reacts to the average while the business runs in the markets.
Sena supplies the consumer signal ahead of the transaction: what people are buying now, what they would switch to, and how large that switch is. The forecasting method stays where it is.
Recorded purchases captured this cycle in each market, so the series starts from now over a closed quarter.
Shoppers state what would move them, so a change in choice reaches the forecast ahead of the sales file.
Every figure names its market, so a regional accuracy number breaks into the countries that produced it.
Financial forecasting is the corporate practice of stating where a period ends: revenue, cost, margin, and cash. It runs on a cycle, refreshing as actuals land, and it feeds the plan ahead of replacing it. Its output is a number the business acts on before the period closes.
The published treatment of this subject is a method inventory, and the top of the search results is a list of techniques. Four appear on almost every list.
Cost and working-capital lines are expressed as a share of revenue and scale with it. Fast, and it inherits whatever error the revenue line carries.
A trailing window smooths noise. It also smooths the turn, so an inflection arrives dampened and late.
Last period times a growth rate. The growth rate is where the whole forecast lives, and it is usually a single figure.
Revenue is modelled against price, shelf reach, and spend. Stronger and limited to drivers already present in the data.
Every one of the four takes a historical series as its input. The technique chosen changes how the past gets projected and leaves the evidence base untouched.
A forecast model is the structure those methods run inside: driver logic, a scenario set, and a link into the ledger. The mechanics are well solved by software. What the model consumes for the revenue line is a demand assumption, and that assumption is where forecast error concentrates.
Risk analysis asks how wrong the forecast could be. Done properly, it needs a measured sensitivity, meaning how far volume moves when price, pack, or availability moves. Modelled from list-price history, that sensitivity reflects the price points already tried. Measured through the stated response, it covers the ones under consideration.
Planning platforms hold the model, the version history, the workflow, and the audit trail, and they hold them well. The forward-looking consumer signal sits beyond every system such a platform connects to, because it exists only where somebody asks a shopper a question.
| Output | What it settles | Where it goes wrong |
|---|---|---|
| Current demand base | What is selling this cycle. | Read from a closed quarter. |
| The turn | Whether choice is changing now. | Detected after the sales file. |
| Switch volume | How many shoppers would move. | Modelled from price distance. |
| Per-market error | Which country drove the miss. | Netted off into a regional average. |
The path read runs on four collected inputs, refreshed every cycle in every market.
| Input | What it answers |
|---|---|
| Receipts | What consumers bought this cycle, at what price. |
| Store captures | Whether the item was there to be bought. |
| Geo-verified photos | The shelf as it stood, dated and placed. |
| Stated preference | The change that shifts the next purchase, and its shopper count. |
The team's own numbers join separately. Ledger actuals, shipment history, price files, and the forecast model connect through 250+ integrations, so a projection meets recorded consumer choice.
Three questions sit outside any ledger, and each one moves the forecast.
Three shipped pages carry the same verb against different subjects. Demand forecasting states the units a market will absorb. Sales forecasting states what the pipeline closes. Revenue forecasting states the money those closes produce at a given price. Sibling financial planning and analysis holds the cycle, and this forecast refreshes inside it.
Real people share what they buy and prefer, under explicit consent. Sena captures it directly at the source, so every figure traces back to where it came from whenever a number comes under question.
Each figure arrives with the market and the capture week attached, which is what allows a forecast override to be defended in a review.
The recorded purchase arrives on a cycle short enough to catch a turn while it is happening.
A commercial view added to the model opens onto the shoppers who support it.
A regional accuracy figure resolves into the markets that produced it.
A forecast built on a closed series can state what happened accurately and states what happens next only by assuming continuity. Sena captures real-world signals in the current cycle, from the shelf photographed this week through to the shopper stating what would move the next purchase.
The purchase and its price are recorded together in the current cycle, so the series starts from now.
Images from real outlets show which items stood on the shelf, so an absent item reads as itself over weak demand.
Signal arrives from the consumer network under explicit consent. The next choice is described by the shopper about to make it.
Ledger actuals, shipment history, price files, and the forecast model connect over 250+ integrations, so a projection meets recorded consumer choice.
Every figure names the market and the week it came from, and a forecast assumption opens onto its shoppers from there.
Every prior forecast cycle with price files and shipment history, scan by scan.
Four teams act on the same forward view, and each one needs a different cut of it before it can commit.
The rolling forecast. Needs a current demand base per market.
The cash position. Needs the revenue timing evidenced.
The scenario set. Needs sensitivity measured over modelled.
The volume signal. Needs the turn caught before the sales file.
The same current base, refreshed against whatever each industry sells through.
The current base per market, with pack switching sized.
Seasonal occasion demand, checked against stock on shelf.
Own-label movement, caught in the cycle it starts.
Product take-up by market, refreshed each cycle.
One mechanism, refreshed each cycle in each market. Each step is documented, which is what carries a forecast override through a review.
The same consumers supply the current purchase and the stated next choice, so a base and its direction arrive attached.
Sena reports the base and its movement per market, so a regional total becomes visible as an average of countries.
A stated change converts to a shopper count, which is the sensitivity a downside case needs.
Most financial forecasts run on a closed series with a commercial override on top, and the consumer sits beyond both. Sena supplies the forward input the model consumes.
| Capability area | Typical setup | Sena |
|---|---|---|
| The evidence base | A closed accounting period. | Purchase recorded this cycle. |
| The turn | Visible once actuals land. | Stated by shoppers before it lands. |
| Sensitivity | Modelled from list-price history. | Measured by stated response. |
| The override | Commercial judgement, unsupported. | Opens onto the shoppers behind it. |
| Shelf availability | Assumed present. | Photographed in the market. |
| Accuracy reporting | One regional percentage. | Decomposed per market. |
| Evidence in a review | A method choice. | Open any figure to its consumers. |
Three forecasting questions where a closed series leaves off.
Read what shoppers are switching to this cycle so a change reaches the forecast ahead of the sales file.
Size how many shoppers a price or pack move would shift, so the downside case carries a measured figure.
Split a regional accuracy figure by country, so the miss is located ahead of being explained.
A measured forward input settles where the period lands. The four decisions below are what a finance team runs around it.
Every finance decision Sena answers, in one place
The cycle this sits in
Holds the plan the forecast refreshes against.
Explore financial planning and analysis 02The forward revenue read
Takes the same forward signal to the revenue line specifically.
Explore revenue predictive analytics 03Delivery against the number
Manages the business to the figure once the forecast is accepted.
Explore financial performance management 04The measurement layer under
Supplies the measurement the forecast and its error are read on.
Explore financial analyticsThe walkthrough takes one revenue line in one region, reports the recorded consumer base for the current cycle in each market inside it, and sizes one switch while the finance team watches.
Name the forecast line and the markets it covers, and we will report the current consumer base under it.