Output vs. Outcomes
Feb 14, 2017 2:20 PM PT
I got a wakeup call from reading "You Need to Manage Digital Projects for Outcomes, Not Outputs," a Harvard Business Review article by Jeff Gothelf and Josh Seiden. The headline says it all.
How many times have we been lulled into complacency over getting a project or a product done but not necessarily well received, because it was someone else's gig? This is a great reminder that the job isn't done simply because we produce something and shove it out the door.
It's also a cue to pay close attention to what our CRM systems tell us.
Is It Done Yet?
If you look at this idea a certain way, it's telling us that our concept of a product has changed -- it actually changed a long time ago, but old rationales persist. A product is no longer the thing itself -- it is the thing plus all the services, processes and procedures we attach. Geoffrey Moore and Regis McKenna before him called it "whole product."
For a long time, whole product was an idea that many businesses could ignore safely. That group is smaller today, but it still exists. The big turning point was the invention of the subscription, or selling a product as a service.
Suddenly, it was a lot harder to push a product out the door and forget about it. Subscription vendors don't make much money on an individual sale and must retain customers for repeat business or they're lost.
Recruiting new customers in this scenario is expensive and can drain the coffers. So getting to outcomes and not simply output for them is critical. What's also critical is that customers get it, and they've been trained to expect subscription-like services from any vendor regardless of what's on offer.
Maneuvering in the Fog
Gothelf and Seiden's article introduces the idea of mission command -- something we might have thought of as taking initiative back in the day. The Prussian army, interestingly, spawned this alternative viewpoint, according to the article.
In essence, it holds that we should rely on individuals to make good decisions to achieve outcomes. This is especially true in war, an area the Prussians excelled at. In the fog of war, the best-laid plans often can be rendered useless by events, so it's important to instill in individuals an understanding of the mission and objectives, while giving great latitude to act in the moment.
If you think the answer is to do more detailed planning, then go back to the fog of war -- it dashes plans with aplomb. Giving the individual latitude in achieving outcomes, therefore, is critical.
Mission command sounds a lot like what it was like to be in sales a few years ago. Communication was primitive; you met with a sales manager once a month and reviewed the pipeline updating as you went along.
The rep had a territory and was responsible for whatever happened in it, so it was incumbent upon the rep to take actions that would benefit the company without checking in with headquarters all the time.
What's new is the emphasis it places on the individual to get things done in spite of AI, machine learning and other nifty new decision-support tools.
Measuring Outcome Success
We're accustomed to staying in our lanes and doing our jobs today, expecting that once a thing is produced that others will take responsibility to get the desired outcome. That's not a terrible division of labor, but it isn't the way things always have been, Gothelf and Seiden pointed out, and the Prussian retrospective experience was useful.
This got me thinking about AI and ML, which lately have invaded the CRM space. It also made me rethink my assertion that these technologies can help us avoid the mistakes humans make when we let our brains short circuit by relying on heuristics rather than actually thinking something through.
It tells me concretely that we need to find the right balance between being freewheeling independent actors in business, and becoming slaves to the information that our systems spit out. It all comes down to what Eric Brynjolfsson and Andrew McAfee wrote about in "The Second Machine Age": that we have to find optimal ways to leverage our machines in a 1+1=3 model.
For me, it all comes down to better listening skills, which start with asking better questions. Open-ended questions about customers' likes, dislikes, and especially customers' feelings related to our products, are the things most likely to tell us how we're doing relative to outcomes and not simply output.
The further we progress in this, the more I see a bifurcation happening. We use a lot of quantitative data to determine success in our output goals, but we need to do better with qualitative data to gauge success in outcomes. We still don't do enough with qualitative data, and if I were an investor, I might look into novel solutions in that area.