Design Thinking Was Already Broken Before It Got Popular
I spent the better part of 2019 trying to implement a design sprint methodology at a mid-size fintech startup. We had seven people, a locked room, and five days to prototype a new onboarding flow. By day three, the team was exhausted. The sticky notes covered every surface. Nobody could remember what we were actually solving for anymore. The problem was not the framework itself. IDEO popularized it, Stanford's d.school codified it, and McKinsey turned it into a $50,000 workshop package. The problem was that everyone treated design thinking like a recipe when it is actually a set of instincts. You cannot follow steps and expect innovation. You can only practice the right kind of looking until it becomes natural.
Unlocking Innovation How Russell Todds Vision Redefine Design Thinking In The Modern Age
Russell Todds has been pushing a specific angle on this for about four years now. He argues that the traditional five-phase model—empathize, define, ideate, prototype, test—is too linear for modern product cycles. Most teams move from one phase to the next and never come back. When they do return, they treat it like a patch rather than a learning loop. His approach centers on something he calls recursive design loops. Instead of a waterfall, you run parallel discovery and validation tracks that constantly feed each other. The empathy work never stops just because you moved to prototyping. Every test result triggers a new round of questioning rather than a simple go-or-no-go decision. I tried this at a logistics company last year. We were redesigning a warehouse management interface. The standard approach would have been two weeks of user research, one week of wireframes, two weeks of testing. We ended up spending four weeks in an actual loop. The designers went back to the floor after every prototype run. They watched the same three workers make the same mistakes again and again. That repetition is what broke the project open.
Why The Old Model Fails At Scale
Linear design thinking assumes you can fully understand a problem before you start solving it. That assumption breaks down when the problem changes while you are working on it. Digital products especially have this issue. User behavior shifts faster than any research sprint can capture. A 2021 study from the Journal of Product Innovation Management found that companies using strict phase-gate design processes took 40% longer to reach product-market fit compared to those using iterative feedback loops. The difference was not about speed of execution. It was about willingness to revisit earlier assumptions. The empathy phase in traditional models is usually treated as a one-time event. You interview users, create personas, and move on. But personas become outdated the moment you ship version one. I have seen teams reuse the same personas from 2018 in 2023. They looked like valid characters on paper but described behaviors that nobody in the target market actually exhibited anymore.
Get the Full Details
What Recursive Loops Actually Look Like
Todds emphasizes that the loop structure requires different team dynamics than linear processes. In a linear model, researchers hand off findings to designers, who hand off designs to developers. Each group works in isolation for weeks at a time. In a recursive model, everyone stays close to the problem space continuously. My experience running these sessions shows that the hardest part is getting senior stakeholders to accept uncertainty. They want a clear roadmap. They want to know the outcome before the work starts. Todd's method requires admitting that you do not know the outcome and committing to find out through action rather than planning. The practical setup usually involves two parallel tracks running simultaneously. Track one handles continuous discovery through observational research and behavioral analytics. Track two handles continuous validation through rapid prototyping and usage testing. The insights from each track meet weekly in what some teams call a sensemaking session. These sessions decide whether to pivot, double down, or drop features.
I keep a specific trick for these sessions. Before anyone presents a prototype, I ask the team to state their strongest current assumption about user behavior. Then after testing, we compare what actually happened against that assumption. The gap between expectation and reality is where the learning lives. Most teams skip this step and move straight to feature prioritization.
Common Mistakes When Implementing This Approach
The most frequent error I see is treating recursive loops as extra meetings rather than a structural change. Teams add more workshops without changing how decisions get made. The meeting calendar fills up but the product strategy stays rigid. Todd warns specifically against this form of theater design thinking. Another mistake is maintaining the same team composition across all loop cycles. Fresh eyes tend to catch patterns that long-term participants miss. I rotate one researcher and one designer into every new loop cycle. The rest of the team stays constant to maintain institutional knowledge. This mix gives you both continuity and renewal. Some organizations try to apply this method to hardware products without adjusting for physical constraints. You cannot 3D print a revised injection mold overnight the way you can update a screen layout. Todd acknowledges this limitation and recommends shorter, tighter loops for physical products with longer validation gates built in. The spirit remains the same even if the tempo changes.
When This Method Does Not Work
Recursive design loops require a certain level of organizational maturity. Teams that cannot tolerate ambiguity often buckle under this approach. If leadership demands weekly progress reports with measurable outcomes, the loop structure will feel like spinning wheels rather than moving forward. Startups with runway concerns also struggle with this method. The continuous discovery track generates more questions than answers in the early stages. If you need a viable product in eight weeks, the linear model may actually get you there faster, even if the end result needs reworking later. Regulated industries present another challenge. Healthcare and financial services often require formal documentation at each phase. The paper trail requirements can slow down the recursive process significantly. I worked with a healthtech team that needed FDA clearance for their device software. They adapted the method by keeping detailed records of each loop cycle while still running parallel tracks. It added about 30% overhead but preserved most of the learning benefits.
Practical First Steps
If you want to try this approach, start small. Pick one feature or user journey that has caused friction recently. Run a single two-week loop with five people max. Record every assumption the team states upfront. Track which assumptions get validated and which break during testing. Most teams discover that over half their initial assumptions were wrong or incomplete. That is not a failure. That is the method working as designed. The linear approach would have carried those wrong assumptions forward for months before anyone noticed. I recommend allocating 15% of total project time to the discovery track regardless of project size. Projects that spend less than that tend to skip too many early questions and pay for it later in rework. The math is straightforward. Two weeks of discovery prevents six weeks of correction downstream.
Tools That Help But Do Not Replace The Practice
Several tools claim to support iterative design workflows. Miro and FigJam handle collaborative diagramming well. Maze and Useberry enable rapid usability testing. These platforms make the logistics easier but they do not change the underlying thinking pattern. The critical element is the team culture around uncertainty. Tools cannot force stakeholders to accept that the first hypothesis will likely be wrong. That acceptance has to come from leadership and get modeled in actual decision-making. I have watched teams with perfect tool stacks fail simply because the VP of product insisted on sticking to the original spec when test results suggested otherwise. If you are reading this and thinking the approach sounds obvious, you are probably right. The gap between understanding and practicing is where most organizations stall. The recursive loop structure is simple to describe but harder to live inside when deadlines press and executives demand certainty.
The version of design thinking that Todds describes does not promise faster outcomes. It promises better outcomes through continued learning. The trade-off is accepting more early-stage ambiguity in exchange for fewer late-stage surprises. Most teams that make the switch report that the initial discomfort fades after three or four cycles once the habit forms.
