Uncaptured

In our earliest pitch deck, we had a graph that showed what we saw (and still see) as our opportunity on the price testing / price optimization side:

The graph visualizes classic supply-demand principles: The lower you price your product, the more buyers you’re likely to have; the higher you price your product, the fewer buyers you’re likely to have.

Intelligems (but also the practice of price and offer testing more generally) has the opportunity to help brands capture more margin. And, the original thinking went, we can generate an immense amount of value for our customers if we’re successful in doing so. Adam and I had seen the story play out during our time in ridesharing, building dynamic pricing and discount algorithms. 

I still believe that, but I now think the opportunity is actually best visualized a bit differently. 

For starters, the slope between price and quantity sold is never this linear.¹ But if we set that aside as a simplification, one of the major callouts is that the graph shows a blended, or  averaged, set of business dynamics.

Since most businesses make offers to their customers, the “list price” isn’t always the price. What the customer pays and is evaluating is impacted by discounts, bundles, subscriptions, etc. Some of these may encourage the customer to buy more units and actually stretch the quantity that can be sold. In that sense, the smaller triangle to the right of quantity sold should also be listed as uncaptured margin. 

So, you might update the graph to look like this: 

In this graph, you are now representing that you are using those offers (discounts, bundles, free shipping, etc) to increase the quantity sold - ie, you are lowering the effective price to create incremental demand. The more you test, the more you can recapture margin using these offers. (You should use Intelligems for this.) And potentially this means you should have your starting price higher.²

This, though, blends each of those offers in the same way the first graph does. It treats those offers as a monolith, which they aren’t.

In fact, the offers are more of a stack.

Today, brands are stacking these offers in a segmented, sometimes even personalized, manner: Sales are run to everyone (no segmentation), but welcome offers are made to new customers (segmentation) and winback offers are emailed to old customers who haven’t bought in a while (segmentation). Certain other discounts are given to CX and allowed to be delivered 1:1 to customers who have a poor experience, like a shipping delay (personalization).

And that stacking means the graph probably looks more like this:

Stacking offers means that brands can meet different customers at different points on the curve, which allows customers to pay the price they’re willing to pay and brands to capture the margin associated with that price (instead of leaving margin unclaimed by giving various subsets of customers better “deals” than were required to win their spend).

The end result of this isn’t a more detailed graph insomuch as it’s a graph where the quantity sold gets stretched to the global maximum (the maximum quantity that can be sold at the unit cost or above) and the price simultaneously gets stretched to the local maximums for each segment (what customers/various customer segments are willing to pay). 

This approach sort of parallels the way Meta and Google has historically won you business: a portion of your advertising spend wins you business below your target CAC, a portion of your advertising spend wins you business at your target CAC and a portion of your advertising spend wins your business above your target CAC.

You don’t see this, though, because it takes place in the auction, and the platforms “wash” that reality by showing you the blended numbers. The result, though, is that both parties win: the platforms are able to increase their inventory (global maximum) and monetize it (local maximum) and brands achieve their stated goals, albeit on a blended basis.

The complexity associated with pulling this off is also a parallel, but that’s what keeps me coming back to this graph.

1   For example, maybe there’s a steady drop off in demand as you go from $70->$80->$90, but then a big drop off “cliff” once you go past $100

2   Topic for a different newsletter, but offers should be managed from a perception perspective. There are tradeoffs to training your customers to always expect discounts