The metrics worth acting on are product-level ones: which items got clicked, which got clicked more than once by the same reader, how long each spread held attention, and where readers stopped turning pages. Impressions, unique visitors, and average session time tell you the catalog was opened. They tell you nothing about what anyone wanted to buy.
The distinction matters because most flipbook dashboards were designed to prove distribution, not to inform merchandising. They answer the question a publisher asks (did people see it) rather than the question a retailer asks (what should go on page three next week), and those are not the same report.
Why Impressions and Average Time Mislead You
A rising impression count usually reflects how hard you pushed the link, not how good the catalog was. Send the same publication to a larger email list and impressions climb while the reading experience stays identical, so the number moves for reasons that have nothing to do with the content you’re trying to evaluate.
Average time on catalog is worse, because it averages two completely different populations. A typical publication has a large group who leave within fifteen seconds and a smaller group who read for four or five minutes, and the mean lands somewhere in the middle describing nobody. Median time, or better still the distribution split into buckets, tells you what average time conceals.
Unique visitors has a similar problem in a retail context. Someone who opens the flyer three times across a week is a stronger buying signal than three people opening it once, and a deduplicated count actively hides that. If your tool can report returning readers separately, that number is worth more than the headline.
The Product-Level Signals That Predict Revenue
Start with the share of readers who click at least one product. That figure typically sits somewhere between 5 and 15 percent depending on category, and it’s the fastest indicator of whether people even realize the catalog is interactive. A number well under that range usually points at hotspot size or a missing visual cue rather than at the assortment.
The more useful cut is the share of clickers who tap three or more products, because multi-click sessions are where baskets get built. In most catalogs a small minority of readers, often under a fifth, accounts for the overwhelming majority of catalog-attributed sales, which is the uneven distribution described by the Pareto principle showing up in reader behaviour, and tracking that group separately changes what you optimize for.
Repeat taps on the same item are the strongest single signal available. When someone clicks a product, keeps reading, and returns to it eight pages later, they have effectively told you they’re deciding. That behavior deserves its own line in the report rather than being counted as two clicks, and it’s the metric worth watching when you’re deciding what to feature on the cover next cycle.
Zoom events belong in the same category. Pinching to enlarge means someone wants detail: a spec, a finish, a price they weren’t sure of. High zoom with low clicks on the same spread is a diagnosis, not a mystery. The hotspot is too small, badly placed, or the page doesn’t look clickable.
Reading Page Depth and Drop-Off Without Fooling Yourself
Attention decays predictably. Most catalogs lose a large share of readers after the first three or four spreads, then hold a long tail who reach the end, and knowing the shape of your own curve is more useful than any single engagement figure. A spread holding readers for 30 seconds when the catalog average is 8 seconds is doing real work.
That gap is where merchandising decisions get made. If page 4 consistently beats page 12 on dwell time but page 12 carries your best margin, the issue is placement rather than product, and moving a range forward in the running order is a clean experiment with feedback inside one cycle.
Position bias needs correcting before you draw conclusions. Bottom-right areas of desktop spreads and anything below the fold on single-page mobile views underperform consistently, so a low click count there is partly a layout artifact. Compare a product against the historical average for its position, not against the catalog average, or you’ll kill a good range on bad evidence.
Absence of clicks is data too, and it’s usually discarded. A product occupying a quarter of a well-read spread that pulls almost nothing over a seven day cycle has told you something no sales report can, because it never got far enough to fail at checkout. Check first that the hotspot isn’t simply broken, which happens more often than anyone admits when linking is done manually under deadline.
What Your Tool Needs to Be Able to Report
None of this works if the analytics only aggregate at publication level. You need clicks attributable to individual products and individual page positions, a way to see returning readers, and an export your team can actually use rather than a screenshot of a dashboard.
That reporting depth is where tools in this category diverge most sharply, and it’s worth reading comparisons such as Flipsnack against FlippingBook with that criterion in mind rather than the feature checklist, because products that look interchangeable for sharing a brochure differ enormously once you need to know which SKU was tapped twice. Ask any vendor to show you a real report naming individual products, not a demo of the dashboard layout.
The other requirement is integration with the rest of your measurement. Catalog links carrying campaign parameters, events flowing into your analytics platform, and product identifiers matching your feed means you can follow a reader from a spread to a purchase. Without shared identifiers, the catalog stays a walled garden and you’ll spend every budget review defending it with vanity numbers.
Budget roughly 20 to 30 minutes per cycle once the setup is right: which spreads held attention, which products drew repeat taps, where the drop-off cliff sat. That’s a review, not a project.
How the Right Metrics Change by Retail Category
Grocery and DIY publish weekly and convert offline, so the meaningful signals are store selector interactions, click-and-collect actions, and offer-level clicks rather than cart adds. Judging those catalogs on ecommerce conversion will always look like failure.
Fashion and beauty produce long sessions with heavy zoom and multiple clicks on a single editorial image, since the reader is buying a look. Furniture and home push further still, where saves, shares, and a return visit three days later carry more predictive weight than any immediate action, because the purchase often involves a second person in the household.
B2B and industrial catalogs invert everything. Cart events barely exist, quote requests and spec downloads are the conversions, and internal search queries inside a 200 page parts document become the richest intent source you have, since every query is a customer telling you what they came for in their own words.
Attribution is the argument that decides whether any of this survives the budget meeting. Someone browses on Sunday evening and buys on Wednesday through a branded search, so last-click models hand the credit elsewhere and quietly build the case for cutting what created the demand. Set the window at seven to fourteen days, agree it with finance before the campaign rather than after, and you’ll be defending a method instead of a number.
