The situation

• The business problem. In-store drove more than 90% of Kroger’s sales, received minority investment, and had no coherent product strategy. We were optimizing the smallest channel and neglecting the largest.
• The org problem. Shopping List, Store Maps, Scan/Bag/Go, and Order Ahead existed as disconnected features with no shared strategy, no common language, and unresolved dependencies between them.
• The customer problem. An early iteration of Store Mode auto-activated on store entry and stripped away everything else. Customer response was openly hostile, and they were right to be.

What I built

Kroger’s instinct was to use in-store moments to pull shoppers out of the physical experience and into digital. I bet the opposite.

Store Mode became an enhancement: opt-in activation, familiar navigation left intact, tools added rather than substituted. And once the mandate was “augment,” five disconnected features could finally be understood as one system.

The arc

One vision, three phases

The vision A digital shopping assistant that augments the store trip instead of pulling you out of it.

Now Clear the noise

Strip out the eCommerce patterns that don’t serve someone standing in an aisle. Make activation a choice.

Next Anchor on the map

Dynamic store maps and real-time inventory become the spine of the in-store experience.

Later Compound the surface

Use the map to reach everything adjacent: scanning, lists, order ahead, routes through the store.

Five disconnected featuresOne portfolio

Named out loud, not buried

Backend readiness Geolocation and privacy compliance Cross-team integration

Three versions

Takeover, toolkit, and a homescreen that knows where you are

Store mode

Version one Replacement

Auto-activates the moment you walk in. Takes over the screen, strips out everything else.

Customers hated it.

Version two Enhancement

Opt-in, familiar navigation intact, in-store tools added rather than substituted.

Augmentation, not replacement.

Version three Geofenced homescreen

Two variants tested head to head, one built around saving money, one around saving time. The standard homescreen stays below the fold.

A learning loop, not a launch.

How I operated

• I shipped the proof before the system. I led feature development for Store Mode, Store Maps, and Shopping List before I ran the In-Store Domain. Those early wins were the first tangible evidence that augmentation drove customer and business value, and they earned me the right to argue for the bigger vision.
• I turned the bet into milestones. I wrote sequenced strategic milestones so partner teams stayed on one path: clear the eCommerce noise, augment the physical trip through dynamic Store Maps, then use Store Maps to reach adjacent features. The trade-off was discipline over speed, and a durable strategy over a series of disconnected releases.
• I named the blockers. The Now / Next / Later roadmap called out backend readiness, geolocation and privacy compliance, and cross-team integration. Naming them made the strategy something other teams could build on, and it’s why the vision outlasted my time on the domain.

What happened

Between 2019 and 2021, the engaged in-store segment outperformed broader app engagement on every dimension we measured.

• Unique customers grew ~23% while native declined ~2%.
• Average order value rose ~25% while native stayed flat.
• Exit rate collapsed from ~10% to near zero.

The more durable outcome was organizational. In-store went from an afterthought to a recognized product domain with its own mission, vision, and portfolio roadmap that later teams built directly on. I moved the argument from “should we invest in in-store?” to “how far can we take this?”

Who owned what

Product and engineering owned the five features I was connecting, which meant every dependency was somebody else’s roadmap. Analytics owned measurement and the segment definitions. Legal and privacy owned geolocation compliance, the hardest gate in the Next phase.

I owned the domain vision, the roadmap and its sequencing, and the Store Mode strategic arc across successive releases. My influence ran up to eCommerce-oriented leadership to fund an augmentation-first strategy, across to the product, engineering, and analytics partners whose features I was connecting, and down to a team of mid and senior designers I built and led against a shared vision.

What I’d do differently

• Formalize the connective tissue earlier. My outcomes lean on Store Mode because it was the most-measured surface. If I’d defined portfolio-level metrics at the outset — plan-to-execute conversion from Shopping List into Store Mode, say — I could have proven the system thesis and not just the feature thesis.
• Separate the opt-in signal from real lift. The engaged-segment numbers are strong, and they’re also self-selected. I’d pair them with a controlled comparison to make the business case bulletproof.