Planning transformation: why most don't deliver, and how to define success, by Vineet Khanna

About Vineet Khanna

Former Senior Vice President and Global Head of Supply Chain at Nestlé. More than 37 years in industry, 25 of them in leadership roles across supply chain, procurement, finance and operational excellence. He now advises companies from strategy through execution and is a regular keynote speaker. Listen to the episode.

Going live is not success

Vineet Khanna has run supply chain for one of the largest food companies in the world and has sat through more planning transformations than most. His verdict is blunt: "Most of the transformations don't deliver business results."

The reason starts with how success is defined. "Going live is not the definition of success. As I often say, not dying is not equal to winning the battle." Teams spend months on workshops, blueprints and change management, and at the end "the only thing you have to show is: we went live without a disaster. I don't think that's how a board will give you money for your next transformation."

This article is his method for doing it differently, plus two opinions that cut against the industry's habits: that orchestration is nothing new, and that technology now comes before process.

Define success in numbers and in capabilities

Business deliverables come in two buckets, Vineet says. "The first bucket is numbers: revenue growth, margin improvement, capital efficiency, ROIC. The second is capabilities." A capability is specific. "I want to be able to grow in e-commerce significantly thanks to technology. Or: today my signal-to-response time is 15 days, I want to reduce it to two days."

Then comes the part most programmes skip. "Keep a hold of the outcome throughout the planning and execution." If the target is 15 days to two, be clear which change in process, which change in technology and which change in the operating model takes you there, and design each of them for that. "Keep a hold of that endpoint throughout the blueprinting, planning, execution and data phase. If you hope that I will do all this investment and at the end of it I will somehow magically reach where I wanted to reach, it will not happen."

The practical consequence is backward planning. "We don't need to do just forward planning. We need to do backward planning from the end point." It also redefines data: "What you feed into your AI models is not just master data of the ERP, you feed knowledge. If I want to bring down my response time from 15 days to two, I need to know what intelligence I need to feed my models in terms of prediction, visibility and simulation."

And when the programme drifts? The point of holding the outcome is that you can see it. "At least you can course-correct."

The science and the art of a transformation

Vineet splits every transformation in two. The science is "the blueprinting, the data side, the project management, change management." The art is "how have you made sure that the key stakeholders in the leadership team are with you."

"No transformation in planning is within supply chain. It is invariably connected with finance, sales and other parts of the business. If those people are not your co-sponsors in some way, these things tend to fail." He lists the usual causes: political dynamics, lack of trust between headquarters and the regions, "people in regions believing that headquarters guys are sitting in their ivory tower and preaching things they don't understand." Co-ownership, communication and managing the bad news during the transition are part of the plan, not an afterthought.

Ben raised the obvious objection: going from 15 days to two is never only a planning-system project; it needs manufacturing, supplier agreements and more. Vineet agrees, and says that is exactly the rigour required. "Co-manufacturers, your customers in some instances and suppliers in most instances. How have you planned those elements? That gets into your planning of 15 days to two days."

Orchestration is the DNA of supply chain, not a new idea

Asked about supply chain orchestration, Vineet starts with a correction. "Supply chain orchestration is a new fancy term, but the whole supply chain came into existence in the mid-90s for orchestration. It was formed when people realised we cannot just magically hope that different functions will connect and align by themselves." The closest older term is systems thinking.

What makes orchestration real is a mindset: "In supply chain there is nothing external. Your suppliers, your co-manufacturers, your customers are a part of supply chain. Unless you consider your suppliers as an integral part of your own supply chain, you're not going to be doing the right things in planning, execution, cost or innovation."

What is new is that technology makes it possible. "Technology today is almost a dream technology. I wish I had it ten years ago when we dealt with many crises." Ten years ago, a what-if in an S&OP meeting meant "we'll come back next week" with a spreadsheet, and asking for 3% instead of 5% meant another week. "Today you can do it at the press of a button, depending on how well you define and connect it in terms of data."

Technology first, then decisions, then process

The industry's standard advice is process first, technology second. Vineet now says the opposite, and he is open that his view changed in the last three or four years.

"Start with the technology capability. You may find that your entire process redesign, which you had done as a ten-step process, is redundant and now you just need a two-step process, because technology is doing the simulation, the optimisation, the connection." Connecting parts of the business no longer needs a process; technology connects them. Periodic processes become unnecessary when technology is continuous. The old L1 to L5 process hierarchies, he says, "are redundant today."

It goes beyond process. "The old operating model, the roles, the organisation structures need to be redefined. They were designed around sequential processes. Today the processes are concurrent." Even a product launch no longer needs the stage gates it used to.

His sequence has one more step at the front. "The fundamental unit at which we need to look at the business is the decision. What decisions do I need to make at the level of strategy, tactics, operations and execution? Start with those decisions, then look at what capabilities technology is giving me to make those decisions, and then come to processes, the operating model and the organisation structure."

If AI isn't paying back, look in the mirror

Vineet's view of what AI can do for supply chain is unreserved. His extreme example is a model solving a century-old mathematics problem. "Do we think that technology cannot make a model of supply chain and solve it? Technology capabilities are evolving very rapidly and they can bring a significant performance improvement in prediction, integration and optimisation."

The gap is organisational. "People are trying to impose technology on the old operating model and the old processes. If you had a six-step decision approval process, you bring in so-called AI, but you still have the same organisation structure and the same six steps, then it's a waste." Hence his line on the ROI debate: "If AI is not giving the ROI, the best place to look is the mirror."

Who is getting it right? Smaller companies, mostly. "They don't have the complex organisation structures and the bureaucracy. Some are starting from a white piece of paper: if I'm starting a small company, will I set up a whole marketing department, or a combination of the right people and the right agents with the right data models?" Large companies have better master data, Ben noted, but Vineet's answer is that the difference is decision speed, not data. One large company he works with is running a slice of its business as an "AI native enterprise" to find out what it would look like.

Your next steps

• Write the definition of success for your current programme in two lines: the numbers (growth, margin, capital) and the capabilities (for example, signal-to-response time from X days to Y). If it says "go live", rewrite it.
• For each capability, name the process change, the technology change and the operating-model change that delivers it, and plan backwards from the end point.
• List the decisions the business has to make at each horizon before you draw a single process step. Then ask which of those decisions technology can now make or simulate continuously.
• Put finance and sales in as co-sponsors, with names, and plan the headquarters-to-regions conversation as deliberately as the data migration.
• Treat suppliers and co-manufacturers as inside the supply chain in your planning design, not as interfaces.
• If your AI pilots show no return, check the approval chain around them before you check the model.

Vineet's advice to anyone building a career in planning is the same advice he gives companies: understand the whole business. "If you have exposure to finance, to sales, to marketing, you're a much stronger supply chain professional. This is the only function which has the word chain in it. The whole DNA of supply chain is about connecting the business end-to-end."