Skip to content
Rank Tracer

Index

All notes

Everything here, in nine sections, including five archive pieces from the site's earlier life kept with their original dates. If you are new to this data, read the fundamentals in order.

Fundamentals

A rank is a position in an ordering, recalculated constantly and weighted by recency. Everything else on this site follows from that.

Reading the data

Charts, comparisons, seasonality and the question of when a movement is real. Most bad rank analysis fails here rather than in the collection.

  • Reading a Rank Chart

    Inverted axis, logarithmic scale, and a small vocabulary of shapes. Getting the presentation right prevents most misreadings before anyone interprets anything.

    Procedure

  • Seasonality in Rank Data

    Catalogue-wide demand swings move everyone's rank at once. Failing to account for it produces confident conclusions about individual products every January.

    Analysis

  • Comparing Two Products Fairly

    A head-to-head comparison is the most common rank analysis and the easiest to get wrong. Six conditions that have to hold before the comparison means anything.

    Procedure

  • When a Change Is Real

    Rank moves constantly with nothing happening. Telling an event from ordinary variation needs a baseline almost nobody establishes.

    Analysis

  • How Long to Observe

    The window has to be long enough to average out noise and short enough that the market has not changed underneath it. Both bounds are real and they conflict.

    Procedure

  • Estimating Units From Rank, With Error Bars

    It can be done defensibly. What that requires, what the realistic error is, and how to present a number that will not embarrass you six months later.

    Procedure

  • A Field Guide to Rank Chart Shapes

    Eight patterns that recur, what each usually means, and the confounder to rule out before believing any of them.

    Reference

Collecting the data

Where the data comes from, what to store alongside it, and why the characteristic failure is a plausible wrong value rather than an error.

  • Where Rank Data Comes From

    Four routes, with different costs, coverage, legal positions and failure modes. Most operations end up with two of them and should understand both.

    Reference

  • Scraping, Terms of Service and the Limits

    The legal position is unsettled and jurisdiction-dependent; the contractual position usually is not. What to weigh before collecting.

    Reference

  • Product Identifiers and the Joins That Break

    ASIN, ISBN, GTIN, SKU and internal codes rarely map one to one. Getting the join wrong produces a dataset that looks right and tracks the wrong thing.

    Reference

  • Storing Rank Data So It Stays Useful

    The dataset is small, the history is irreplaceable, and most of the value comes from context fields people do not capture at the start.

    Procedure

  • Building Your Own Tracking, Realistically

    The collection is a weekend. The maintenance is indefinite, and it is what determines whether the dataset is trustworthy in year two.

    Analysis

  • Gaps, Nulls and Silent Data Failures

    The characteristic failure in rank collection is not an error but a plausible wrong value. Four monitors that catch it, and the manual check that catches the rest.

    Procedure

Marketplace mechanics

Reviews, price, advertising, availability, listing structure and shared sellers. The mechanisms that produce the movements you are trying to read.

Applications

Competitive tracking, campaign measurement, demand signals and reporting — with a worked example taken from question to recommendation.

Pitfalls

Correlation traps, survivorship bias, manipulated data, vendor claims, and the questions where the honest answer is that this data will not do it.

Beyond one marketplace

What transfers to other marketplaces and other ranking systems, and what has to be rediscovered locally.

  • Other Marketplaces and Their Ranking Systems

    Rank systems differ in what they measure, how visible they are and how they can be tracked. What transfers from one marketplace to another and what does not.

    Reference

  • Search Position Is Not Sales Rank

    One is a published scalar; the other is a personalised, query-dependent distribution. Measuring the second properly requires a different method entirely.

    Analysis

  • App Store Charts and Other Ranking Systems

    The same reasoning applies to app charts, streaming charts and book bestseller lists, with different mechanics that change what the number means.

    Reference

Reference

Terms defined once, including the several that are used inconsistently across the industry.

  • Glossary

    Terms used across this site, defined once. Where usage differs between vendors and marketplaces, the definition here is the one these documents assume.

    Reference

Archive, 2010–2012

Five pieces from this site's earlier life as a rank tracking service, reproduced with their original dates. The figures are historical; they are kept as examples of the method applied at the time.

53 notes in total