How we’re rethinking our approach to OKRs and goal-setting

Fab Dolan / June 2020 / Data & Measurement

All Google marketing plans include a solid measurement strategy, anchored by short- and long-term OKRs (objectives, and key results). In any year not numbered 2020, this requirement has kept our marketing campaigns rooted in business objectives, and it’s a part of our culture on which we pride ourselves.

I like to say that inside Google marketing, you can either be a scientific creative or a creative scientist, but you have to be one. Yet three months into the most uncertain and volatile crisis of our time, it strikes me that our marketing rigour, in this context, has become a liability.

Establishing objectives, forecasting results, measuring attribution — each of these benefit from high degrees of certainty. But during this crisis, certainty about anything is hard to find. Our models are built on historical data, yet we’ve never seen media or buying habits like this. Our financial assumptions are developed using industry benchmarks, yet standard CPMs and CTRs are shifting in real-time.

OKRs are great when you can (to some degree) predict the future, but they’re less instructive to your team in times of uncertainty.

It’s now impossible to feel 100% confident in our plans. As a result, we’ve changed our approach to goal-setting and audited our measurement and research behaviours. Google’s marketing measurement evangelistAvinash Kaushik did a great job of establishing some do’s and don'ts on specific tactics.

At a more philosophical level, here’s how I’m seeing our OKR process change across Google marketing.

We’re managing for principles — not targets

Principles can act as superior guidance to your team, over specific objectives and key results. One of the most challenging elements of this uncertain environment is just how rapidly things change. We’ve had to let go of good ideas on multiple occasions because in the week it took to draft a brief, establish goals and measurement framework, and refine the idea, the ground had already shifted beneath our feet. In other cases, we are left outright guessing about what targets to set for our fall campaigns.

This can take a toll on your team. To solve this, we’re increasingly trying to establish principles, not hard OKRs. OKRs are great when you can (to some degree) predict the future, but they’re less instructive to your team in times of uncertainty.

Principles like “develop moonshot programs to help main street retail small businesses survive and recover from this crisis” offer better guidance at the moment. In this example, teams can look at every creative idea, budget decision, partnership or media buy, and choose the option that instinctually will drive more support for retail small businesses.

Operating with principles instead of hard targets is not a license to avoid strategic thinking altogether.

We’re checking assumptions and theories, to understand sensitivity

Operating with principles instead of hard targets is not a license to avoid strategic thinking altogether. Drafting core principles is more difficult than the status quo of forecasting and measuring all the standard marketing performance indicators. But even with good principles, we need some semblance of what good and great looks like. We also need to believe that specific initiatives will be in the ballpark, even if we don’t know where they’ll land on the field.

One way we’re doing this is by articulating our assumptions and hypotheses in advance, and debating the strength of our hypotheses as a team. For example, while forecasting outcomes for a hardware launch, we might ask ourselves to declare assumptions such as: when the economy rebounds, what base demand for the category might be, or what percentage of retail will stay online. Of course, only hindsight will answer these questions in full, but these early discussions will expose any uncertainties around specific parts of our plan and will pull out from the group all possible viewpoints from the data and experience we have today.

We’re sculpting instead of designing

You can constantly refine your approach to objectives and key results as you go. As you move forward in time, you’ll learn more. You’ll get real-world data and results and see developments from your product team. You’ll get feedback from sales. All of these things might adjust how you feel about your original assumptions and theories. Like a sculptor, you’ll be able to take a broader, less-defined idea and begin to add definition to it as you go.

It’s important to recognize how much uncertainty you face for a particular decision, and to calibrate your OKR approach to match.

Allow new data to speak to you. If you’re using digital platforms, iterative target setting is easier. Digital platforms can capture a lot of information by default and allow you to work backwards. As long as your measurement hygiene is strong (and even if it isn’t, digital platforms have strong residual memory) you can put something out in the world that aligns to your basic principles. Then, ask deeper questions like which demographics it resonated with, what types of specific actions resulted, and how it changed overall user sentiment.

The greater the uncertainty, the more we rely on thoughtful debate about principles and assumptions. The more we debate and learn, the more we make sense of the world and can better add detailed objectives and key results to our principles. It’s important to recognize how much uncertainty you face for a particular decision, and to calibrate your OKR approach to match.

A framework to help guide marketing strategies now, and in the future