What Competitive Intelligence Means in Supplements

"Keep an eye on the competition" is probably the most frequently given and worst defined advice in e-commerce. In practice it means, for most people, an occasional tour of a few known competitors' pages — done when there's time, meaning rarely, meaning exactly the rhythm at which every important move arrives a month late.

Competitive intelligence is something else: a system with defined questions, defined sources, and a defined rhythm. This first chapter lays the foundation — what can be known, what can't, and how the work splits between machine and human — because the rest of the guide installs the system piece by piece on this foundation.

What you'll be able to do after this chapter

• name the four families of public signals every honest monitoring effort is built on;
• say explicitly what you do NOT see about competitors — and reject conclusions that pretend to see it;
• understand the system's architecture: what you delegate to the machine, what remains your judgment.

The principle

About a supplement competitor, four families of things can be known from public sources — and RavenBI gathers them all under one roof:

1. What they promote — their active ads, with their signals: age, estimated reach, variants, launch rhythm. Not their whole marketing, but the part they expose publicly through paid media — the signal of the priorities they're testing, not of their results.
2. What they sell — the products in the catalog, the ingredients they push, the Amazon presence with its evolution visible there.
3. How they sell — their funnels: the pages the ads lead to, the offers, the trust elements, the change history.
4. Where they're heading — the derivatives: accelerating or cooling, raising or lowering their visible launch rhythm, entering new categories or disappearing from the tracked signals.

What can NOT be known, however good the tool: sales, costs, profitability, conversions. Even spend is only estimated — RavenBI computes it from the reach Meta publishes times market CPMs and labels it explicitly as a rough estimate. An honest system works with what exists and names the rest: "competitor X launched 40 ads this month" is a fact; "competitor X is doing well" is a guess dressed up as a conclusion.

The second half of the foundation is the division of labor. A human can't continuously track hundreds of brands — and doesn't have to: tracking is machine work. What the machine can't do is judgment: whether a competitor's move changes any of YOUR decisions. Hence the architecture of the whole system, which you'll find again in every chapter: the machine observes the covered signals and filters; you judge, on a fixed rhythm, only what passed the filter.

The workflow in RavenBI

1. Take the orientation tour of the four families, so you know where each one lives: Ad Library (what they promote), Products and Amazon BI (what they sell), Funnel Radar (how they sell), Brands (where they're heading — the page that synthesizes them all per brand).
2. Open the page of a competitor you know well and compare what you see with what you knew: the What changed (30 days) section shows what they launched NEW — ads, products, ingredients, pages. The gap between what you knew and what you see is exactly the size of the hole this guide plugs.
3. Note the brand type labels: Advertorial, Publisher, Retailer — not everything advertising supplements is a competing brand; the radar separates them, and you'll read a direct competitor differently from an affiliate pushing someone else's product.
4. Write down the short list of YOUR monitoring questions — "who's attacking my category?", "what are my three main competitors doing?", "is anyone new appearing on my problem?". The following chapters turn them, one by one, into system pieces.

How to read the signals

• Public fact (launched, stopped, changed) = the system's raw material — recorded without interpretation;
• Computed derivative (accelerating, cooling) = a direction signal — more valuable than the isolated fact, but still not a performance verdict;
• Labeled estimate (estimated spend) = an order of magnitude for comparisons, not a number to report onward;
• Anything that sounds like "it's working for them, do what they do" = a step over the border — public data shows the competitor's effort, not their result.

The typical mistakes

• monitoring only the competitors you know — the ones who hit you in a year are usually the ones you don't know today; the system must see the market, not your list;
• confusing activity with performance — a competitor launching a lot may be burning money; what you see is their bet, not their winnings;
• reading estimates as measurements — estimated spend is a comparison compass, not accounting;
• building vigilance instead of a system — any routine that depends on "remembering to look" dies in two weeks; that's why the following chapters install pieces that work on their own.

Exercise

Write the honest answer to three questions: when did you last find out about an important competitor move, how late did you find out, and what would the delay have cost you if the move targeted you directly? Keep the sheet — at the end of the guide, the same three questions will have different answers, and the difference between them is the system's value.

Every tracking system starts with the same decision: what exactly you track. Not "the market" — the market is too big; your entities, chosen deliberately. Next chapter: the Watchlist.