# The number won’t move. Nobody has time to find out why. You own this company. Mid-hold. Great brand, capable team, and a margin that will not budge. This is what would happen if you brought us in, told in pictures. Illustrative composite: the company and numbers are representative. The method is exactly how we work. 87% of the first stream’s hours would add no value: waiting, chasing, redoing +34% throughput within 90 days 30 days to the first signal the number would move Honest by design Representative engagement. Real method, illustrative company (a mid-market consumer-goods distributor) and numbers. No client is depicted. The unease ## You can feel the drag. You can’t point at it. The company, an illustrative consumer-goods distributor, sells more every quarter. The margin does not follow. The gap between those two lines is where your best people are drowning. **Decisions reopen.** The same question returns to the same three people. **Work comes back.** Context dies in every handoff, so it gets done twice. **Heroics are the process.** The best people rescue the flow by hand. Again. ### The two lines that don’t agree Illustrative chart: revenue climbs quarter over quarter while margin stays flat. The widening gap is the operational drag. Growth keeps buying more work. The work keeps eating the margin. The call ## One conversation. One place to start. Ninety minutes of listening. Six workflows on the table, and one to start: the one where the drag is loudest. Six candidate workflows on a board. Order-to-delivery is marked start here. Finance close is marked healthy and left alone. Wholesale motion is marked not yet: something is there, but it is not the first move. The rest are deliberately untouched. Marketing ops Customer service Order-to-delivery Wholesale motion Finance close Product dev Small enough to finish. **Real enough to matter.** Act one · See ## Two weeks inside the actual work. A Bottega workshop would map two core workflows end to end, sitting with the people who carry them: order by order, handoff by handoff. What the map would show ## 87% of the work would not be the work. 87% **Non-value-add:** hours no customer would ever pay for. The most capable people in the building would be the most buried in them. **Measured in the room, not from a dashboard.** Two weeks sitting with the people who carry the orders: timing real work with a stopwatch, writing down every wait, every chase, every bounce. ### One order, door to door A value stream map of the distributor's order-to-delivery workflow. Five steps hold about four hours of real work. Between them sit twenty-three hours of waiting in queues, and rework loops send orders back to earlier steps: one in three orders is re-picked, and bad addresses caught late go back to re-keying. Of thirty-one elapsed hours, roughly eighty-seven percent add no value. **How the 87% is measured:** every elapsed hour goes into one of two buckets. Teal is real work, the kind a customer would pay for. Coral is waiting, chasing, and redoing. Twenty-seven of the thirty-one hours are coral. The same hours as one bar: 87 percent non-value-add, 13 percent real work. Where the wasted hours went: waiting on handoffs 24 percent, rework and corrections 21 percent, re-keying between systems 17 percent, chasing status 15 percent, approvals that change nothing 10 percent. The work that matters: 13 percent. Waiting on handoffs 24% Rework and corrections 21% Re-keying between systems 17% Chasing status 15% Approvals that change nothing 10% The work that matters 13% **Not a people problem.** A process that grew faster than anyone had time to redesign, illustrated with representative proportions. The hard call ## We’d find two problems. We’d fix one, and tell you why. The workshop would surface two bleeds Fix now ### Order-to-delivery - Evidence: clear, mapped, measured - Team: ready and asking for it - Payback: fast, and it funds the next move Not yet ### Wholesale sales motion - Tangled into three other systems - Mid-migration, owned by nobody - Slow, uncertain payback at high cost Real money leaking. Still the wrong first move. Honest, **even when it costs us the bigger invoice.** Telling you not to fix the wholesale motion yet would be the most valuable call of the engagement. We’d make it every time. Act two · Shape ## Research shows that workflow redesign, not the model, is what makes AI pay off. Bolt AI onto a tangled flow and you get a faster tangle. We redesign the flow first, so the tools have something worth speeding up. ### Where AI value actually comes from BCG's 10-20-70 rule as a stacked bar: 10 percent of the value is algorithms, 20 percent is tech and data, 70 percent is people and process. Paraphrased: McKinsey’s State of AI tested 25 factors; redesigning workflows around AI was the single biggest predictor of bottom-line impact. BCG’s 10-20-70 rule corroborates. ### Redesign the flow, then add AI. The order-to-delivery flow before redesign: eleven steps, six of them waste. Before · the flow we’d find **11 steps · 6 pure waste** · every coral box is where the margin goes The redesigned flow: five steps. Agents carry the re-keying, status, and checks; people keep the one judgment call. After · the flow we’d shape **5 steps · 0 re-keys · 1 decision** · AI only where it pays, people on the judgment ### Each improvement funds the next. Three sequenced moves, each funding the next, each with a real OKR. Move 1 #### Order-to-delivery The loudest drag, fixed first. OKR: cycle time, cut in half →funds Move 2 #### Inventory sync Paid for by move one’s recovered hours. OKR: stockout rate, floor it →funds Move 3 #### Returns loop By now the program pays for itself. OKR: refund cycle, days not weeks **No leap of faith.** Each move pays for the next, and each has one number everyone can watch move. The result ## 31 hours would become about 11. The margin would finally move. Same workflow we’d map in act one, measured the same way: door to door, order by order. Here is what would change, and where the money shows up. ### One order, door to door: the rematch Scorecard for the order-to-delivery flow, before and after the redesign. Door to door: 31 hours before, about 11 hours after. Steps in the flow: 11 before, 5 after. Orders re-picked: 1 in 3 before, near zero after. Throughput: baseline before, plus 34 percent within 90 days after. **More orders per person.** The same team would ship about a third more orders without adding a single hire, so the cost to fulfill each order drops. **Fewer fires, fewer refunds.** Orders would stop bouncing between steps, so rushed shipments and make-good refunds mostly disappear. **Cash shows up sooner.** An order that sits in queues for a day and a half would leave the same day, and the cash for it arrives sooner too. ### Where the margin gain comes from A simple margin bridge for the order-to-delivery flow. Starting from the margin before, two gains stack on top: more orders per person, then fewer rushed shipments and refunds, reaching a visibly higher margin after. Proportions are illustrative. Directional by design: the shape of the gain, not a client's P&L. Same team, a third more orders. **More of every sale survives to the bottom line.** Act three · Build ## Our own tools would build it, from the inside. One coherent operating change. Not four products to manage. The workshop Sits inside the work and shapes the new flow around the way the team already operates. The skill builder Captures what your experts know and turns it into skills the whole team can lean on. The pipeline Carries every build from design to shipped, reviewed and verified at every gate. [Korium](korium.html) Underneath all three, quietly learning everything the work touches: decisions, reasons, lessons. The memory layer **AI only where it earns its place.** People keep the decisions that need them. The exit ## What your key people know becomes shared institutional knowledge. Before: the company's know-how locked inside three individual people. After: one person at the center of a connected shared memory, with the company's memories flowing in around them. Korium carries it across. The company is one resignation away from losing what it knows. **It doesn’t have to be.** The know-how lives in two or three heads today. It would move into Korium, your shared memory: the taste, the decisions, the reasons behind them. People can take a vacation, or a better offer, and the company still knows what it knows. The team keeps it. So does whoever you eventually sell to. The payback ## First signal in 30 days. Step change by 90. Then it compounds. A payback curve over 120 days: first measurable signal near day 30, a plus 34 percent throughput step change by day 90, then compounding gains after the engagement ends. ~30 days **First signal.** Less rework, fewer reopened decisions, hours coming back. ~90 days **Step change.** Roughly 34% more throughput from the same team. After **Compounding.** Recovered revenue funds the next move. The gains hold after we leave. The point ## Your best people get their real jobs back. The same people, doing the work that actually needs them. A company that keeps what it learns, and is worth more because of it. One more time, honestly ## The story is illustrative. **The method is not.** Same eyes, same tools, same willingness to tell you the truth, pointed at your portfolio company. Starting with one place, done right.