Instacart is experimenting with new grocery models, and some of them may not work out. Increasingly, I think Instacart "is what it is" and will have an extremely difficult time developing new models.

Its recent ButcherBox partnership tells a story. A subscription box service for grass-fed beef launched on Instacart in major markets. This is not going to work.

• I can't search on Instacart for grass-fed beef and find ButcherBox. Because they are only offering "boxes of beef" and not individual "beef" items. This means Instacart's catalog model does not understand bundles.

Why pick a box partner if your algorithm doesn't understand bundles? How could it possibly succeed?

• Even within the ButcherBox store on Instacart, searching for "grass-fed beef" returns no results. Eh? I've said before that Instacart COULD in theory be helpful to launch new DTC brands, but the experience cannot treat each of these brands as its own independent store.

At the top level, Instacart is a browse experience, not a search one. You search at the retailers you already know, except in specialty situations.

Instacart has gotten big on the backs of retailers. It needs to detach from that model and take huge inventory/supply chain risks to take its next step. Is it willing to?