The 5 greatest pitfalls of Design Sprints (and the solution)
Everyone’s done them, everyone’s had at least one that didn’t go well. Realistically, a lot more than one.
Design sprints, much like design thinking, have served many industries well for decades, but things have shifted, and their value has diminished significantly. However, in our collective urgency to build better products and find solutions as rapidly as possible, Design sprints have become a default choice to identify a solution and move forward. At Gyroscope, we believe that AI presents a new chapter in solution discovery and definition, phasing out the traditional design sprint. With that in mind, let’s walk through five of the greatest pitfalls brought on by Design Sprints and how AI can circumvent them all.
Caveats
Design Sprints are meant to be more of a framework than a methodology. Unfortunately, we see/hear examples that are more process-driven and rigid in nature.
Our critique is more of a one-week GV-style Design Sprint. They work great in certain contexts, but require modifications to be effective outside of their intended context.
AI is not the only way to solve the problems listed below. AI is the strongest and most time-efficient way to modernize a process that holds much value in all creative endeavors.
1. Time
The antiquated problem
Ok well… yeah this one feels obvious no? Five days to define a problem statement clearly, identify a viable solution, test, prototype and then leave a plan for the client to continue working from creates intentional urgency. Such a compressed timeline for any form of creative work means that more often than not, these projects are destined to fail.
How often have you seen teams return to the old ways of doing things after a completed design sprint?
The modern solution
AI drives value when applied correctly, whether it’s basic automation or Generative AI, as a force multiplier, finding solutions or answering questions faster. What we believe that means for a design sprint is not necessarily to output more or maximise the findings in the traditional 5-day sprint but actually to do more and gain better insights in less than five days.
2. Poor Problem definitions
The antiquated problem
For those who have taken part in design sprints, it can often feel like one camp is eager to believe that the execution of a design sprint is a magic bullet to solve a problem, while either stakeholders are not bought in, sometimes even going so far as to withhold specific data that would be pertinent to the problem that is trying to be solved. Inevitably, this creates an error carried forward scenario where the design sprint is destined to fail before it even begins.
The solution presented by AI
I’m about to talk about LLMs; please hold your groans for now, I promise its worth it. It’s completely understandable that one might roll their eyes when the AI solution to any problem seems to boil down to an LLM. Still, there is a reason for that: so many issues we’re looking to solve with AI today focus on interpersonal communication or connectivity. LLMs and their baseline conversational models mean they’re instrumental in scenarios like problem definitions, where it’s unclear how to find common ground between over-eager stakeholders and non-believers who are cynical about the potential solutions.
An LLM model can rapidly process the data and accelerate problem definition, allowing everyone to agree on it.
3. Engagement in process
The antiquated problem
Design sprints depend on participants' engagement. Participation is fundamental to how they work, but buyers have noticed a steady decrease in the ROI of design sprints for the past few years, partially attributed to the decline in participant engagement. There are a few reasons: if you’ve been around a company long enough, you may have been subjected to far too many Design Sprints, and the mere thought of taking part in one brings on fatigue. Not everyone is creative, and that’s ok, but putting them in a room and forcing them to take on the designer role for 5 days isn’t just daunting; it’s exhausting. It’s also frankly unfair to expect non-designers to produce solutions at the rapid-fire pace that designers are accustomed to.
The modern solution
One of the most significant advantages of leveraging generative AI models has been enabling those who are unable to visually communicate through traditional methods like sketching, 3D modeling, or other visual mediums start to express the visuals from inside their heads out into the world for others to see what they’re thinking.
AI levels the playing field for creatives and non-creatives, allowing them to communicate solutions and concepts they believe to be effective without fear of being misunderstood.
4. Prototypes that probably don’t do a lot
The antiquated problem
Design sprints are super intense; anyone who has taken part in one can tell you about the exhaustion you feel on the final day. What they may fail to mention is that a large part of fatigue comes from the volume of prototyping that happens during the sprint. While prototyping is necessary, the speed at which they’re executed in a design sprint often leads to unrealistic solutions or expected concepts. Neither option is ideal.
The modern solution
Not all Design Sprint problems fall on the shoulders of the participants, facilitators are also required to ensure they can channel what their participants are doing into something meaningful. Technologies like vizcom, Photoshop AI, and to a lesser extent, Midjourney enable facilitators to capture images of prototypes and rapidly uplevel them with polished-looking renders.
In the context of a hardware effort, a higher-fidelity digital output helps others understand the design intent alongside a lower-fidelity physical prototype.
The resulting images help fill the gaps for those who may struggle to know how a cruder prototype works and create excitement for participants on the potential future for the job done in their design sprint.
5. Incomplete Testing
The antiquated problem
Design sprints are inclusive but testing needs to consider a range of potential users. Shutting down or pausing a team for five days to participate in a design sprint is challenging enough. Getting the budget and paperwork together to involve external test users for early-stage prototypes can also be another, often larger, challenge. The net result? Incomplete testing, which in a best-case scenario results in prototypes that will be iterated on further after the sprint is complete. Unfortunately, incomplete testing results in a bias of results from the testing done during the sprint, creating insufficient data that are carried forward through product development, inevitably resulting in revisions and changes later.
The modern solution
User testing is critical in the development of any product. We have seen many well-intended efforts from design sprints go sideways when data points of one/verbatim comments are taken out of context and used as justification for a less-than-ideal outcome. To solve that particular problem, we have developed virtual personas built using available demographic and psychographic data to create archetypical personas for potential users. These personas allow workshop participants to ask questions about the efficacy of their solution to get valuable feedback as early as possible. When concepts exit the workshopping phase, they can be subject to higher scrutiny with actual users specifically recruited for the research. Are you interested in learning more about how our AI-enabled design sprints, aka Guided Discovery, can benefit your organization? Reach out to us here