If we are talking about size variations (selling a pair of socks in different sizes but the same price), I would start with one PGN or EPBA set holding ALL KWs and ALL variations. After a while (30 to 60 days), simply pause the variations with the lowest CTR.
As an example let’s say we sell spatulas with 3 different materials (wood, bamboo, silicon).
Start the same way.
Additionally:
For a high volume keyword like “spatula“, obviously, you would want to show which spatula material has the highest CTR (let’s say “silicon spatula“) to get the initial click and let customers that click through choose whatever variation they prefer to buy once on the listing.
However, there are keywords that would be specific to one of the variations and you would only want that spatula to show (e.g. “wood spatula” or “bamboo spatula“).
The END GOAL would be to have:
The “wood” and “bamboo” PGNs will have ONLY dedicated KWs and you would negate the “wood” and “bamboo” words (Negative Phrase) out of the MAIN PGN, and negate the general “spatula” words (Negative Exact) out of the ancillaries.
This will give you 100% segmentation, perfect search term isolation, and control on where the traffic is sent, in order to both keep your offer hyper-relevant to buyer search and keep the data and automation nice and tidy.
Once more, In my opinion, the easiest way to get to this point is to start with ONE PGN or EPBA set holding ALL KWs and ALL variations, and do 2 things:
– Identify what the “MAIN” variation is
– Identify what the ancillary Keywords are
Go about splitting them only after you have all this data.
Think about the 80/20 Pareto Rule:
– If the Silicon Spatula is driving 80% of searches and 80% of sales, why bother to start 3 PGNs or EPBA sets up front?
– If you do not know what variation will be the 80/20, what better way to ask buyers to vote with their clicks (and start with All in ONE PGN or EPBA set)?