← All guides

BoardTempo workshops · 6/8

Three Pins, three outcomes: an analytics decision workshop

Три пина, три разных результата: учимся принимать решение по цифрам · BoardTempo
Original BoardTempo editorial illustration created with AI.

One Pin has ten thousand impressions. Another has three thousand. The first one wins, right? Before commissioning ten more images just like it, look at the next column. Then the one after that. The loudest person in the room is not always the person moving the project forward.

This is a decision workshop, not a tour of every analytics menu. We’ll review three Pins pointing to one pottery studio article and decide what to try next. Every number below is invented for teaching. These are not BoardTempo results, client statistics or performance benchmarks.

Put the evidence on one table

Imagine an article about choosing a handmade mug as a gift. Pin A shows an atmospheric breakfast. Pin B compares two handle shapes. Pin C offers a gift-selection checklist. For this exercise, assume the same observation window and organic distribution. In real reporting, verify those conditions rather than quietly assuming them.

PinImpressionsSavesOutbound clicksTagged website sessions
A: breakfast scene10,0003005040
B: comfortable handle3,000309072
C: choosing a gift1,000202015

Impressions count appearances on screen, not unique people. Outbound clicks take someone to a destination outside Pinterest; they are not the same as opening a Pin. The sessions in this exercise come from a separate, fictional website analytics report. We are comparing two measurement systems, not two labels for the same event.

Calculate before you tell the story

Outbound click rate for A is 50 divided by 10,000, multiplied by 100: 0.5%. B is 3%; C is 2%. These figures help compare our fictional rows. They are not a universal scorecard for good and bad Pinterest performance.

The save rates are 3% for A, 1% for B and 2% for C. Now the picture is more useful. A attracts saves but relatively few visits. B has fewer impressions, yet more outbound clicks and website sessions in absolute terms. C sits between them on rates, with fewer observations overall.

A does not necessarily have the “wrong audience”. People may be collecting breakfast styling ideas. That is an understandable use of the image; it simply differs from our current goal of attracting readers to a mug-selection guide. If our objective were visual awareness, we would ask a different question.

Name what the table cannot tell you

We have no purchases, qualified enquiries, revenue or acquisition costs. So “B makes the most money” is not a conclusion supported by this table. Even an extremely confident presentation cannot conjure the missing column.

We also need publication dates, seasonality and distribution conditions. Some of those have been simplified for teaching. A real report should record the date range, data source, filters and extraction date. Today’s incomplete data should not be presented as though it were a finished week.

Most importantly, this is not a randomized experiment. Pinterest may show each Pin to different audiences under different circumstances. A rate difference suggests something worth investigating; it does not establish that the handle photograph caused the result.

Why didn’t fifty clicks become fifty sessions?

A click and a website session are different events. Page loading, analytics configuration, consent choices, repeat visits and redirects can affect what gets measured. The gap alone does not identify the cause.

Start with the destination URL. Do its tracking parameters survive the redirects? Does measurement run under the conditions it is intended to? Are the reporting dates and time zones aligned? Test without collecting unnecessary personal information. If measurement is unavailable, “sessions unverified” is a more useful statement than false precision.

Our UTM guide explains the basic link labelling. For this workshop, the important outcome is traceability: you should be able to connect each report row to the actual Pin and destination, not merely to a similar-looking image.

Give each row a next action

For A, review the promise. If the headline only describes a beautiful breakfast, saves for styling inspiration are not surprising. Try a genuinely different concept that answers a real mug-buying question. Changing the border colour does not change the reader’s reason to click.

For B, keep the practical-comparison hypothesis and test another relevant feature, such as capacity. One encouraging row is not a permanent creative formula. The next example still needs to answer a question covered by the destination article.

For C, check whether the promised checklist is easy to find. Then gather more comparable observations. Twenty clicks are twenty clicks, not a licence to declare the strategy proven. If visitors cannot find the checklist, fix that problem before blaming the Pin’s colours.

Sources

Share this article on Telegram

Share something useful

Share on Telegram

Let’s talk about it

What could help your project? Share your experience or ask a question.

The discussion will appear here after this article is posted in our Telegram channel and comments are enabled. You can visit the channel meanwhile. Comments are public; you are not automatically subscribed.

Telegram loads only when you click. Telegram may ask you to log in to comment. If the discussion is not available, use the channel link above.