I guess I don’t really mind that people generally regard me as being opposed to salesperson autonomy (and, of course, performance pay). Antagonism is certainly more newsworthy than conformity! However, my position is not quite this simple, and a deeper understanding of the basic principle at stake here is more useful than the sound bite. If you’ve viewed our multi-media presentation or read our whitepaper (you can get both free on our website), you’ll remember I present a simple cloud (conflict diagram). I suggest that managers typically equivocate between two approaches to the engagement of salespeople. On the one hand, management wants salespeople to be autonomous agents (think, sub-contractors). On the other, management wants salespeople to be team members (traditional employees). I argue that these two approaches are mutually exclusive: in reality a person marches to the beat of his own drum OR subordinates to the team’s drumbeat. I also argue that the case for autonomy is not as strong as is commonly assumed — providing management with the option to engage salespeople as true team members. Now the operative word here is ‘option’. My argument is not that autonomy — and everything that goes with it (including performance pay) — is necessarily bad. Rather, I argue that, what is destructive is management’s attempt to equivocate between two mutually-exclusive alternatives. (More on the consequences of equivocation in a moment.)

The case for autonomy

Imagine that you are an executive at Hewlett Packard. You have two divisions, one manufactures inexpensive computers for domestic customers (moms and dads). The other division engineers custom solutions for enterprise customers. Think about how you distribute your inexpensive computers. Do you have salespeople knocking on doors in the suburbs? Of course not, you distribute them via Harvey Norman (think, Circuit City if you’ve never heard of Harvey Norman). Now, Harvey Norman is an autonomous agent. They bring enormous value to the table for you: massive stores (and car parks), a captive audience, logistical capability and so on. Let’s call this value ‘infrastructure’. However, when we engage with Harvey Norman — an autonomous agent — we must sacrifice some of the benefits we’d enjoy if we managed sales internally. We can’t walk the floor at Harvey Norman and issue directives to sales clerks. It’s also unlikely that we can collect data at the most granular level like we can within our own organisation. So, there’s a trade off. And with less information and less control it’s inevitable that we will not be able to integrate (synchronise) production as tightly with our sales agent as we would if we owned the entire channel. And, as I hope you’ve already guessed, we compensate for this imperfect integration by maintaining inventory. In other words, we build domestic computers to replenish stock, as opposed to building to fulfil direct customer orders (like Dell does). As is always the case, poorer integration equals larger inventories. In this case (domestic computers) we’re quite comfortable with this trade-off. Ultimate sales are made by an autonomous agent — and we pay this agent on a piece-rate (performance pay) — and we’re better off, in-spite of the need to carry inventories.

The case for team membership (integration)

Let’s consider our other division. This division sells custom technology solutions to enterprise customers. These solutions — by virtue of their being custom — are engineered to order. You can see the problem already, can’t you? If these solutions are engineered to order, they cannot possibly be sold from inventory. If we can’t build inventories of stock to buffer imperfect integration between production and sales, then we’d better strive for perfect integration. Even though perfection is unachievable, it still shouldn’t be hard to see that there’s simply no comparison between the degree of integration we can achieve if we own the entire distribution channel as opposed to what’s possible when sales are handled by autonomous agents. In fact, it may well be that tight integration between sales and operations is the most effective way for you to gain advantage over your competitors. (If you operate in a pure engineer-to-order environment, please go back and read that sentence again!) So, in this case, it doesn’t make sense for sales to be outsourced to Harvey Norman or to any other variation on the ‘autonomous agent’ theme. In this division (enterprise solutions), sales and production must both march to the beat of the one drum. However, when we take sales ‘in-house’ we must recognise that we sacrifice the infrastructure that a Harvey Norman brings to the table. And, obviously, considering our market, we’re not particularly uncomfortable with this trade-off.

Give and take

In summary, then, when we distribute through autonomous agents we give up control but gain infrastructure. When we own the entire channel we give up infrastructure in favour of greater control.


I can almost hear you thinking, a typical salesperson is hardly comparable with a retail giant like Harvey Norman. And I’m glad you’re thinking that: it’s exactly my point! When we (management) equivocate between engaging salespeople as autonomous agents and team members, we end up with the worst of both worlds: limited control and negligible infrastructure. So, my position is not that salesperson autonomy is necessarily wrong. It’s that equivocation between autonomy and team membership is. You need to consider your organisation, and its market, and determine which approach to salesperson engagement is the more appropriate. After all, you can’t eat your cake and have it too.


In many organisations it’s quite okay to have a mix of the two approaches within the one channel. For example, while it doesn’t make sense for the final link in Hewlett Packard’s distribution chain to be fully integrated, it’s likely that HP has a team of salespeople who call on Harvey Norman stores (channel managers). These channel managers should absolutely not be autonomous agents. Even though Harvey Norman is a good example, an autonomous sales agent can make sense even if they don’t bring quite the same scale to the table. For example, if we have a manufacturer client, we will often turn capital-city salespeople into true team members (as per our Sales Process Engineering model) and encourage our client to engage autonomous agents in regional areas — who may well be individuals who sell a range of non-competitive lines.