Is it ever okay to have a proxy product owner, and can a product owner role be filled by a consultancy, someone in IT, or some other non-employee? For those familiar with my perspective, you might be surprised to hear that my answer is yes. It’s okay to be an arm’s-length product owner, but the trade-off always comes at a cost or with attendant risk. That’s the topic of this segment.
There are two core categories of arm’s-length product owners: the proxy owner and what I’ll call the discretionary owner.
A proxy owner is someone serving the product owner role but without the concomitant decision power. This person is likely very competent within the domain space, and s/he understands the vision of the product. In most cases, a decision made by this person will be in sync with what the actual owner would make, or, even if the decision is not exactly what the true product owner would make, s/he is nonetheless okay with it.
Proxy owners are either lower-level business functionaries or proxy tokens from the IT staff or a consultancy as scrum team members. The risk in proxy owners is that they’ll make decisions that will need to be amended or undone, or there might be a delay in acquiring the necessary direction. This may retard progress and adversely impact schedule, budget, or both. It may also have a negative effect on team cohesion or morale. If you acknowledge these risks and are okay with them, a proxy owner is fine. If the risk manifests and you need to blame something, the first suspect is the decision to staff this position with a proxy.
I see a discretionary owner as parallel to a stockbroker with discretionary power. For those unfamiliar with this part of investment management, there are ostensibly two types of stockbrokers. Most stockbrokers are non-discretionary brokers. They can only execute a trade with your permission. Like a generic product owner, they understand your goals and objectives and can develop an approach with you. But unless you call (by whatever means) or you respond to a call and give your explicit consent, no action can be taken. The aforementioned proxy owner is a nice analogy to the non-discretionary broker. In the case of proxy owners, the delay might adversely impact the project. In the case of a broker, it might cause you to lose value or the opportunity cost of value.
One gives a discretionary broker power to execute trades without the owner’s consent. If an investment meets some pre-defined criteria, s/he can buy or sell immediately. This removes the time-delay risk related to non-discretionary accounts (or proxy owners). But it’s not without risk. It creates principle-agency risk. Perhaps s/he makes a trade that s/he feels is in your objective space, but you yourself would not have made. If this investment happens to work in your favour, you may say no harm no foul. But if it is a losing investment, there may be hard feelings beyond the financial loss.
There is a slight difference between the discretionary broker and the discretionary owner. The broker will make the decision, and in an instant it will be done. The equivalent to a product feature function being ‘in production’. A discretionary product owner will likely have prioritised some item in the backlog and assigned it to a sprint. As these are probably at least two-week cycles, the worst case is that someone might prematurely commence design or development time, thereby using time that might have been better spent, but this feature function is unlikely to see light of day in production. That is unless the product owner is truly absent.
Again, if this arrangement fails to produce optimal results, blame the decision to employ this model rather than giving Agile a black mark.
In summary, it is okay to have absent or proxy product owners. Just realise that you are risking suboptimal execution and outcomes. If you’d prefer to be an optimist, you might tell yourself that perhaps the proxy or discretionary owner would make better decisions. If that’s the case, you’ve got different challenges.