Voters 1
Sold / impressed ratio ! critical
- Open
- Subscribe
 |
X D |
I'm a Data Scientist and would rank the products by geometric mean (multiplication of chained probability transitions [0. ... 1.] in customer journey from the first touch up to the sale). This is exact problem that we are trying to solve here in real life when building the shop. By the reality, step by step: ads -> journey+discovery -> sale.
And I can feel that the formula for profit starting with first impression in advertising up to the sale is something like:
avg_volume_of_auditory_to_impress * ( (p1 * ... * p__of_n_steps_in_customer_journey) * (avg_sell_price - avg_cost_price) - avg_cost_per_1_impression )
Let's decompose it for clarity:
CrossProbability of user's journey in the funnel: (p1 * ... * p_of_n_steps_in_customer_journey)
let just simplify it to ratio: sales/impressions
sales/impressions * target_profit_per_sale - cost_per_1_impression
this is our profit per 1 impression
... * volume_of_auditory_to_impress
we scale it by auditory (how many people will see our ad)
simplified:
( sales/impressions * target_profit_per_sale - cost_per_1_impression ) * impressions
So from the "success" perspective we have 2 the most important parameters to measure to name some product a "winning":
what is the "sales/impressions" ratio
what is the "target_profit_per_sale" we can apply on the market
@AutoDS , so we need "Trending products" page to be adjusted.
1. Critical. Please add a Filter on ratio Sold / Impressions. We have right now "Interaction rate" but this is not what is needed.
2. Current selectors
2.1. "Impressions" - now it has "more than X" behaviour. We need "Less than" behavior.
2.2. But better to make a more flexible selectors for every field! For each filter it would be great to have 2 fields of "greater than" and "less than" with a field for numerical input