Sena dates the crossing from purchase, ahead of a category total.
Detection is a timing problem dressed as an analysis problem. The methods for confirming a trend are sound, and they all require enough evidence to be confident, which by definition arrives after the movement is well underway. The commercial value sits in the gap between a shift being detectable and a shift being obvious, and most detection practice operates at the far end of that gap.
The evidence required to be certain accumulates with every cycle that passes. A team that waits for certainty gets it, and gets it at the point where acting carries the least advantage. The judgment is a trade between being early and being right, and treating it as a purely analytical question buries the trade where a commercial decision belongs on the table.
Detection rules get written for a category and applied across all of them. A threshold appropriate for a slow-moving staple reads flat in a fast category and floods the list in a volatile one. The rule stays in place because revisiting it needs a judgment, and the output looks orderly either way.
Shifts start somewhere, and detection usually runs on aggregates. A movement that has crossed decisively in two countries is diluted below the threshold by nine that have yet to move, and the aggregate crosses a threshold long after the earliest market did. The countries that move first are frequently the smallest contributors to the total.
Sena captures purchase at a frequency that makes small movements visible in every market separately. The crossing point can therefore be dated where it happened, rather than inferred from a total that moved later.
The capture window a movement crossed in, established rather than estimated.
A crossing date per country, so the earliest is visible on its own.
Sensitivity set against how fast the category itself moves.
Early trend detection is the decision that a movement has become substantial enough to treat as a trend. It sits between a signal, which might resolve either way, and a named trend, which has been sized.
Trend detection covers any method for identifying that a measure has begun a sustained move, rather than fluctuating. The whole difficulty is separating those two states early, since a sustained move and a fluctuation look identical for the first few observations. Every technique for doing it trades false positives against delay, and choosing the balance is a commercial decision rather than a technical one.
Social trend detection watches posting volume, hashtags, and sentiment for early rises. It is genuinely fast, and it detects conversation, which precedes purchase inconsistently and by a variable margin. It produces early warnings with a high false-positive rate. Used to nominate candidates that purchase evidence then confirms, it plays to what it is good at.
| Error | What it costs |
|---|---|
| Too early | Engineering spent on a movement that faded. |
| Too late | Arriving in a market a competitor already holds. |
| Too sensitive | Attention spread across a list of everything. |
| Too strict | A real shift confirmed once it is common. |
Most detection practice is tuned to avoid the first and pays heavily for the fourth. A false start costs one project. Confirming a shift after competitors have built against it costs a position, and the second is recoverable far more slowly than the first.
Three properties separate a real movement from a fluctuation before the size makes it obvious. Persistence across capture cycles, presence in more than one market, and a stated reason that matches the observed switch. A movement holding all three at a small size is stronger evidence than a larger one-cycle spike in a single country, and the conventional threshold, which reads size alone, ranks them the other way round.
A crossing is an event in a country, rather than an event in a category. A shift that crossed in Vietnam and Indonesia two cycles ago and remains flat in nine other markets is detected and actionable, and an aggregate view reports it as below threshold. The commercial move is to build for the two and watch the nine, which requires the crossing dates to be held separately.
Sensitivity appropriate for a staple is wrong for a fast-moving format, and one rule applied across a portfolio is wrong somewhere.
Passing a threshold on one large cycle and passing it on four consecutive small ones carry different weight, and a rule reading level treats them alike.
An aggregate threshold reports a category, and the build decision is about a market.
| Output | What it settles | Where it goes wrong |
|---|---|---|
| Crossing date | Which cycle it became real | Estimated after the fact |
| First market | Where it began | Lost inside an aggregate |
| Persistence | How many cycles it held | Judged on a single spike |
| Corroboration | Whether the reason matches | Absent from a volume rule |
The first two make a detection actionable, and the last two make it credible. A movement dated to a cycle in a named market, holding across three captures with a stated reason behind it, is something a team can fund. The same movement described as an aggregate that recently passed a threshold is something a team can argue about.
| Window | Where an onset appears |
|---|---|
| Receipts | The cycle units began moving between formats. |
| Store captures | The first appearance of the format on shelf. |
| Geo-verified photos | Where and when, placed to the outlet. |
| Stated preference | Whether the reason given matches the switch. |
The watch list, detection rules, and alert thresholds keep the owner they have. 250+ integrations bring them alongside the purchase read, so a rule-based alert and the units behind it share one row.
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 capture cycle arrives dated and placed, so a crossing can be pinned to a window and a country.
Figures in this exchange are illustrative.
A dated window per movement, per country.
How many captures a movement has stayed above the threshold.
Whether what consumers said matches what they bought.
A reporting cycle can confirm a movement once its own reporting period closes. When the movement began, it can say only by looking backwards through periods, which is the one direction a closed period cannot be re-observed in.
A reporting period closes and becomes one figure. Every cycle behind this page is retained on its own, in its own market, which is what allows a crossing to be dated to the window it happened in.
A recorded purchase places the cycle units began moving in, which is what a crossing date rests on.
Shelf photography catches a format arriving, giving the movement a supply-side date alongside the demand one.
Signal arrives from the consumer network under explicit consent. Whether a new choice has settled is answered by the person making it.
Threshold definitions, escalation rules, and the candidate list attach over 250+ integrations.
An onset date carries the capture window and the outlets that established it.
Every cycle is retained separately, which is what makes persistence measurable at all.
Own the detection call. Need persistence rather than a single spike.
Own the timing. Need the earliest market named.
Own the category rules. Need sensitivity set per category.
Own the local crossing. Need their own country’s date.
The cycle a format began taking units from another.
Occasion shifts dated before they reach the category total.
Movement to a new delivery format, caught at onset.
First movers on a product feature, dated.
One watch, held across cycles until a movement clears a threshold set for that category, beside the detection rules already running. Every crossing keeps its date and its country, so a call can be reviewed later.
Purchases and captures are read per market, so each country stands on its own movement rather than contributing to one total.
The cycle in which a movement first cleared its threshold is recorded, giving the crossing a fixed point rather than an estimate.
Persistence, market count, and stated reason are read together, which is what separates a crossing from a spike.
Most detection waits until a movement is large enough to be certain about, by which point it is also public. Sena moves the evidence earlier.
| Capability area | Typical setup | Sena |
|---|---|---|
| Basis | Aggregate volume passing a line. | Purchase captured per market. |
| Timing | Once the total moves. | When the first market moves. |
| Geography | One category-wide call. | A crossing date per country. |
| Persistence | Judged on one period. | Counted across capture cycles. |
| Sensitivity | One rule for every category. | Set against the category’s own pace. |
| Corroboration | Volume alone. | Stated reason beside the switch. |
| Evidence in a review | A threshold report. | Photographs, dated and placed. |
Three timing questions a threshold rule answers poorly.
Locate the country a movement crossed in first, ahead of the category total.
See global consumer intelligence →Check how many consecutive cycles a movement has held above its threshold.
See competitive shelf intelligence →Use a crossing date in one country to plan the timing in the ones behind it.
See market entry and expansion intelligence →Every innovation decision Sena answers, in one place
Before the question arises
Carries the reads at the stage where a crossing is still far off.
Explore market signal intelligence → IN2Sizing what crossed
Takes a detected movement and measures how much of the category it holds.
Explore consumer trend intelligence → IN7The series behind the call
Supplies the line a crossing gets judged against, cycle by cycle.
Explore trend analysis → IN1What a crossing funds
Holds the build agenda a detected movement then has to compete on.
Explore innovation intelligence →The session takes a movement the team is watching, finds the cycle it first cleared threshold in each market, counts how many captures it has held since, and checks the stated reason against the observed switch.
Bring the movement and the markets you watch it in.