concidea
Get in touch
How Pythia is built

May 2026Stage 2 · Explain

What did the newsletter do?

Impact per send and promo codes, conservatively attributed.

Sketch: newsletter on day T, effect on the following days and a ranking
Sketch

Newsletters go out regularly. Which of them really made a difference last year? And does a discount code bring more than a newsletter without one? The open rate doesn’t answer that. The question is what was sold afterwards – and whether it was more than you would have expected anyway.

Every newsletter becomes an event

Pythia imports marketing’s send list. Every newsletter becomes an event on its send date, with discount and code as properties. It then appears on the timeline like any other event and gets its own impact card.

The reference is what matters: a newsletter about garden furniture should be measured against garden furniture, not against total revenue. Pythia assigns every newsletter to a range based on its subject line – or deliberately to “global” when it is aimed at everyone, such as for promotion weeks or the new year. None is left unassigned.

Impact in a short window

Pythia measures a newsletter’s impact in a short window after sending, three days by default, adjustable from one to 28. It measures actuals against expectation, plus a control period beforehand, channel share and code redemptions.

Over a period this produces a ranking of all newsletters with a summary: how far above expectation were they typically? How many were above it at all? And how do newsletters with and without a code differ?

The ranking is an observation, not a recommendation. It shows what was different after a send – not what to do next time.

Correcting what the machine gets wrong

Automatic assignment is sometimes wrong. So every row has a pencil: reference, code, discount and note can be changed. The correction survives every new import and applies from then on to all newsletters on the same topic – including the reminder that follows a few days later.

Promo codes as a view of their own

Discount codes become a filter and a table of their own. Pythia groups them into promotions – the same code within a continuous period – and suggests a readable name containing topic, period and discount. Anyone who prefers to name a promotion themselves can do so; their name is kept.

Next month: every number explains itself – explain mode and the rule register.

Talk to us about Pythia

YOUR ENQUIRY

What can we help you with?

Topic

We only use your details to handle your enquiry. More in our privacy policy.