Why fast feedback cycles are a strategic asset
Speed up feedback cycles to drive better decisions and create self-optimizing teams
Feedback cycles are an under valued component of business strategy.
A few books - notably The Lean Startup - focus on the value of building a culture that embraces fast feedback cycles when starting a company. But few companies think about the speed of feedback cycles when making strategic decisions (like deciding which products to launch or where to allocate resources).
Rapid feedback cycles have a big strategic benefit: faster learning. They also unlock a new set of organizational capabilities, including:
A low reliance on great “top-down” decisions
Less time spent on planning and decision making
A more autonomous, self-optimizing workforce
Access to bottoms-up “big ideas” that lead to big impact
In this post, I’ll lay out why fast feedback cycles matter, why your team should consider them when making strategic decisions, and how to increase the speed of feedback cycles in your existing business.
Let’s dive in!
Slow and Fast: Ford and Facebook
I’ll illustrate the impact of fast feedback cycles using two examples: Ford and Facebook.
Ford is the quintessential example of a company with very slow feedback cycles.
Ford’s most important point of feedback is the same as any company: the profit they generate from the products they sell. Unfortunately, Ford has to make product decisions long before their products are sold.
Cars are designed and engineered years before they hit the market. Capital is put into production lines decades before most cars roll off the line. Ford may have to wait years to get real feedback on the decisions they make.
Facebook is the quintessential example of a company with very fast feedback cycles.
Facebook also primarily cares about the profitability of their products, but unlike Ford, Facebook receives near real-time feedback on every product or feature they ship.
Facebook’s primary source of profit driver is advertising. The amount of ads Facebook can serve to users is driven by the amount of time users spend on the platform. By measuring the impact of every product change on the amount of time that users spend on their platform, Facebook is able to measure the impact of that change on revenue. Since they have very high margin business, revenue is directly tied to profits.
This allows Facebook to behave very differently from Ford.
Ford has to make a few big decisions correctly: which cars to launch, which plants to build. Because these decisions are very important, Ford spends a lot of time and resources trying to get them right.
Junior people complete analyses to inform Ford’s big decisions, such as production forecast, customer surveys, and focus groups, but junior people don’t make big decisions themselves. Executives make all the big calls.
Facebook still has to make some big decisions (looking at you Metaverse) but most important product decisions aren’t made by executives. Instead, Facebook lets individual engineers and PMs try new interesting ideas and see what happens.
Since Facebook can measure the impact of every change on the output they care about in real time, they can give employees the freedom to try things. If an employee ships something that doesn’t work, they can just roll it back. Inversely, if an employee tries something and it works really well, it can be expanded to a billion end users overnight.
Fast feedback cycles are the foundation of Facebook’s culture of autonomy and freedom. Slow feedback cycles are the root of Ford’s culture of top-down decision making.
Of course, there are structural reasons why Ford is Ford and Facebook is Facebook: Ford sells a very expensive hardware product with high fixed costs, Facebook makes a free consumer software product.
But Ford could be more like Facebook, and they could gain a lot from that. Tesla is proof that this is possible in the automotive business.
How Tesla got faster feedback cycles
Tesla isn’t Facebook. If you’ve read Walter Isaacson’s Elon Musk biography, you know that Tesla has a very top-down decision making culture.
Still, Tesla has made a few strategic decisions that give them much faster feedback cycles than Ford. One example is their decision to sell their vehicles directly to end customers, rather than going through dealerships.
In the automotive industry, dealers buy cars from automakers months in advance. Dealers have to make purchasing decisions without knowing how strong demand for the vehicle will be. Once the cars arrive at a dealership, they’ll either sell quickly or sit on a lot for months.
Because of this arrangement, automakers get predictable revenue. They know that they’ll be able to sell most of the cars they make to dealerships as soon as they’re manufactured.
Unfortunately, it also takes a long time to feel the pain of bad decisions - automakers only feel the pain when they have to give sales credits to dealers to move vehicles off the lots, to make way for the next model year.
Tesla doesn’t operate with dealers. There are a lot of reasons for this, including a desire to keep the margins captured by dealers and Elon Musk’s obsessive desire for control. But working without dealers gives them a much faster feedback loop. This gives them a bunch of advantages over Ford’s model:
Better signal: Tesla knows right away if a new model is or isn’t selling, and can track outcomes in real time
Fast follow-ups: They don’t have to wait for the year to end, so they can start making changes to configurations or pricing right away
Allowance for more, smaller decisions: Tesla can do things like adjust pricing in real time, and make multiple, small adjustments
Easier decisions: Since Tesla can easily adjust their pricing, each individual change has less weight; they don’t have to get it right the first time.
Tesla’s direct sales strategy was a big strategic choice with many other drivers & implications. But getting faster feedback loops was a major benefit of this choice.
We’ve established that faster feedback cycles unlock new, valuable capabilities. We’ve also seen that it’s possible for companies to create faster feedback cycles by making big strategic decisions.
So what does this mean for you?
1: Feedback speed should be a criteria you use to decide which product, project, or initiative to launch
Let’s say you’re a SaaS company. You need to expand, and you’re choosing between two new products:
A new product that will primarily used by enterprise customers
A new product that will primarily be used by self-service users
Even if both products are successful, it will take a lot longer for the first product to see adoption and show results. In the interim, the team will continue making decisions without immediate feedback.
This makes any incremental product decisions harder, since the team has less information to use. Leadership will probably worry about if it was the right decision, so they’ll ask for more check-ins because there isn’t concrete data. All of this adds a lot of work and inertia!
If they launched the self-service product, these problems would mostly go away. They’d have dozens of new companies using the product within the first month who they could interview and run tests on. Each decision would have immediate feedback.
Doesn’t that sound like a better choice?
Maybe, maybe not. If the TAM for the enterprise product is 100X the TAM for the self-service product and no one else serves the market, you should probably work on the enterprise product.
Still, the speed of feedback cycles should be a consideration as you make the decision.
2: Leaders should work to increase the speed of feedback cycles
Even if you’re pursuing a project that typically has slow feedback cycles, you can do a lot to increase your speed of learning.
Let’s run with the “enterprise product” example. By default, let’s assume you’d study competitor products, create a product roadmap, build all the features on the roadmap, then launch the product to enterprise customers. It will take 6 months to do all this, at which point you’ll get the first set of feedback from real customers.
What can you do to increase the speed of this feedback cycle?
In general, three components of a feedback cycle.
When thinking about about how to increase the speed of feedback cycles, most people focus on the speed of decision-making and action. However, these are really hard to change they are largely determined by culture. Culture is very sticky.
If you want to increase speed here, you have to make long-term, concerted effort and set a very strong example. This includes things like:
Setting very clear expectations for timelines on follow-ups (hours, not weeks)
Increasing your team’s tolerance for mistakes, to allow for more output
Working at night or on weekends to quickly deliver work items you’ve committed to
Ruthlessly eliminating unnecessary process or bureaucracy
All of these are incredibly important, but they aren’t going to drive change quickly. If you’re Ford, it will take years to drive a meaningful cultural shift.
However, you can increase the speed at which you collect feedback. People respond very well to feedback, because it makes decisions easier. This means decisions are made faster. It won’t necessarily change how fast action is taken, but it will accelerate your cycles.
So how do you increase the speed at which you gather feedback?
When asked this question, people immediately think of ways to gather proxies for real feedback - not gather real feedback faster. Think focus groups, customer interviews, surveys, etc.
Ford is an expert in gathering proxies for real feedback. Their budget for surveys, focus groups, and forecasts is probably in the billions.
Unfortunately, gathering proxies for real feedback doesn’t accelerate the feedback cycle. Proxies are not conclusive evidence that something does or doesn’t work for customers, so the outputs are subject for debate. More importantly, collecting proxy data drives extra work that’s unrelated to building the product or delivering it to customers. This is a waste of time.
So what is “real feedback”, and how can you collect it faster?
Real feedback is concrete data from an end user who is using a product in a “real” environment. Some examples:
Data on usage of a beta product (e.g., activation, retention, time spent, etc.)
Questions / requests from customers who have access to the product
Recordings of user sessions
Content generated from users (e.g., tweets about the product)
Gathering real feedback means you have to show early versions of products to customers. However, there are some low-risk ways to do that:
Build with your customer: Find a customer that wants the product and is willing to work with you as it’s built. Give them access to the product immediately, so they are actually using it. Incorporate their feedback at every step of the journey.
Ship a beta version: Instead of waiting for the “final” version, publish a beta version as soon as possible.
Ship to a small group of customers: Give a small group of customers access to a rough version of the product.
Sell it quickly: Don’t wait until it’s “ready” to let people spend money on it. If you really don’t think it’s ready, offer pre-orders.
Collect real-time data: Establish metrics to track every measure of a product’s success. If the metrics you have take too long to provide signal, find metrics higher up in the funnel.
You’ll notice that Ford could do some version of all of these. Here’s what it could look like:
Build with your customer: Give test versions of a new vehicle to customers for free, and let them drive the vehicles and share feedback
Ship a beta version: Sell a limited number of test vehicles to dealers
Ship to a small group of customers: Produce a small number of vehicles and only sell them in one state
Sell it quickly: Allow refundable pre-orders for every vehicle
Collect real-time data: Gather daily data from dealerships on the sale of every vehicle
There are probably some cost and regulatory constraints that it hard for Ford to sell “beta” versions of vehicles or put test vehicles in the hands of real drivers. Still, it seems like it would be worth paying the legal fees to avoid a big flop.
If Ford can do these things, your company can too.
These sorts of changes are relatively easy and quick to make. If you have a team developing a product, all you have to do is expose that product to customers in some way.
Most people are extremely hungry for feedback on their work, and they will use it if you give it to them. Real feedback cuts through bullshit arguments.
What type of company do you want to run?
The more you create fast feedback cycles, the more autonomy you can give to the people who work for you.
Fast feedback cycles let employees “self-optimize” since they can tie the work they do directly to an outcome. This means they get more independence, more freedom, and more room to try crazy ideas.
This makes life easier for managers, since they don’t have to drive centralized processes or micromanage individual decisions.
It can also drive huge impact for a company when one employee’s crazy idea turns into a runaway success. Facebook Marketplace started out as a hackathon project; now it’s core product.
If you’re the CEO of Facebook, you ride a tidal wave of incremental improvements. Rather than stressing about earnings, you get to spend your time working on problems you care about (like the Metaverse).
If you’re the CEO of Ford, you have to get every decision right. Instead of getting to design a sweet new car, you spend your time agonizing over huge investments.
Personally, I’d way rather be the CEO of Facebook than Ford.
Most of us won’t be the CEO of anything, but we can make our teams feel more like Facebook. Speeding up feedback cycles is an easy place to start.