Hi Andrzej,

The problem with every problem, pun intended, is that your brain insists on seeing it as a straight line. Something is in a certain state today, so somewhere back on the timeline there must have been a cause that put it there. Find the cause, intervene on the line, get the change you wanted.

We look at systems the same way. Take a three person sales team with one hyperperformer and two reps who keep missing target. The obvious first move is to swap out the two laggards. Obvious, and usually wrong, unless you look at that team the way Donella Meadows would look at it.

A sales team does not live in a vacuum. It sits inside a web of dependencies, internal ones between the reps themselves, between sales and marketing, between sales and finance, and external ones between your team and the prospects it is supposed to convert.

In my first job, one of the reps stood out from the rest of us. He beat target regularly, and by enough that the people who were actually hitting their numbers looked pale next to him. The sales director tried to fix the gap by asking our star to train everyone else. The frustration when nothing improved after those sessions only lifted when it turned out our star had a deal running with the head of marketing, the person responsible for lead generation. She fed him the best qualified, fattest leads, and more of them than anyone else got. He returned the favour by splitting his inflated bonus with her under the table.

What complex mechanisms produce, and that includes human organisations and the sub-organisations inside them, depends far less on the individual elements than on the relationships between those elements.

Elements of a system are small systems themselves, wired into a bigger one by relationships

Meadows pointed out that the goals of individual elements can produce behaviour in the whole that nobody intended. When the rules, the incentives and the flow of information do not tie local decisions to the goal of the organisation, people can genuinely improve their own results at the expense of the company's. What finally comes out is decided by the connections just as much as by the parts.

For a company, the absence of a deliberately designed discount policy does not mean the absence of rules. What governs instead is whatever is already there: the existing bonus scheme, the habits, the informal understanding of who can approve what. Often different rules in different people's heads.

So when you look at your discounting, the question worth asking is: what behaviour does our system make easy and reward, and which consequences never reach the people making the decisions?

Before we get into the mechanics, it is worth being clear about what you are fighting for here. Across 50+ pricing projects for sales-led B2B companies, we have identified discount tightening potential ranging from 5% to 50%. That is the full spread, and the tails are real in both directions. The median sits around 17% revenue uplift. If your sales-led motion is anything like the ones we have worked in, that is roughly what is sitting on the table while everyone argues about whether the two laggards should be replaced.

What is a discounting system made of?

A discount policy describes part of the rules. A discounting system also includes the people, the tools, the way customers react, and the information about what your decisions actually produced. Even well written rules will behave differently than intended if the rest of the machinery pushes behaviour somewhere else.

The map below translates a selection of Meadows' concepts into discounting. It is a practical application of her approach, not her own model of commercial policy.

These mechanisms all run at the same time. Picture recurring end of quarter discounts. Some customers learn to wait for the better offer. Those delayed decisions increase the pressure on the sales number. That pressure pushes the company into the next round of discounts, which hardens the expectation. You get a reinforcing loop. That is a mechanism to go and check in your data and in conversations with your customers, not a claim about your company.

You can also design a balancing loop. When deal outcomes drift from the goal you set, the information reaches the people who own the offers. They change the terms on the next set of deals and check whether the drift is shrinking. A dashboard becomes part of a loop like that at the moment somebody uses it to make a decision and then to judge that decision. Wiring consequences back to the people who caused them is a core part of the Meadows approach.

A good policy allows a discount when the discount helps hit a goal you consciously chose. A lower average discount does not prove anything on its own: you may have shed profitable contracts on the way. You have to judge the effect on the whole business, and you have to allow for the time it takes for that effect to show up.

Next steps

Go and check whether your discounting system supports the goal you say it has. What comes out of this exercise is a short verdict on four mechanisms, plus the single biggest contradiction or knowledge gap you found. You can run it with whoever owns your commercial terms.

1. Write down the goal you will judge decisions against. Pick one type of discount and finish the sentence: "This discount should help us achieve _, over _, while maintaining ___." For example, winning new customers in a given segment at an agreed minimum deal profitability. Confirm that goal with the process owner, and write down any answer that differs from yours.

2. Pull five real discount decisions in that scope. Include one where the later consequences are already visible, and include refusals if you have any. For each one, record the terms, the justification, who approved it, and how anyone checked it against the goal. Mark missing data as missing.

3. Check the incentives of the people deciding. Put the goal next to their KPIs, their bonuses and what their managers expect of them. Can they be rewarded for a discount that breaks the agreed criteria? If they can, point at the specific rule or the specific case from the five you pulled.

4. Check whether the rules support the stated goal. Do they let someone grant a justified discount, and do they catch the deals that fail the criteria? Who allows exceptions, and on what basis? Write down which rules actually did something in the cases you reviewed.

5. Trace the information before the decision and after it. Does the approver get what they need to judge the deal against the goal? Who receives the later consequences, including the ones that only surface at renewal? Who changes the next offer or the rule on the back of that? Note both the connections that work and the ones that are missing.

6. Write down the verdict on four mechanisms. Four rows: incentives, rules, information at the point of decision, and feedback. Give each one a rating: aligned / contradictory / unknown, and a fact that justifies it. Leave the unknowns as unknowns. An absence of information is not evidence of alignment.

Do let me know what you find.

And seriously, if you run this and you are not sure how to read the results, reply to this email. We can go through it together, free of charge.

Stay deliberate,
Maciej

Feel free to reply to this email if you find it interesting. Also, if you're not in my LinkedIn network - let's connect!

Maciej Orczykowski
General Manager & Partner