Starting a Multi-Brand Marketplace With a Design Sprint
How KI challengers — A KI group company — used a fully digital Design Sprint approach to start a new marketplace. This article was written by Pedro Cardoso 👏
Our final product was a multi-brand digital marketplace in the MENA Region focused on fashion and accessories.
This story is not so much about the success of the marketplace but about the process we used to get started in our early concepts and designs, why we used it, and how it helped us achieve our early goals.
So, who are we? We’re KI challengers. We bring specialists in Product Management, Product Design, Strategy, and Engineering together to help others create their own companies and digital products. Along the way we learn and create some of our own as well.
As with all interesting stories, this story has a beginning, a middle and… an open end (like many current TV shows, you’ll need to wait for next season to know what happens next).
Context
The owner of the multi-brand marketplace is a family-owned business established in the late 19th century, with retail franchise businesses across the world. With steady growth, it’s one of the leading franchise enterprises in the world, with more than 3000 store locations, from flagship stores, to shopping malls, and even your local coffee shop.
At the time, they’d been growing their presence online but had yet to explore building a multi-brand digital marketplace where customers could have a similar experience as in one of their malls.
As with many businesses, the 2020 pandemic pushed them to quickly evolve their online strategy as people couldn’t shop in retail shops during lockdowns. This was the challenge and the beginning of this story.
Our client was in conversations with other agencies but KI challengers was able to prove that we could pull this off well and quickly. So, in June 2020 we assembled the right team to start this endeavor.
People
We had a great, diverse, international, cross-functional team from the start coming from both KI challengers and the client. Most of the experienced key roles necessary were in place, available and eager to participate in the beginning of this discovery. This story is all about discovering the best path to follow and how important it is to have the right team involved from day one.
A key point here is that everyone was part of the same process, same meetings, same concerns, same findings, same outcomes, same failures, same solutions, and same accomplishments, thus co-owners of everything.
It was both fascinating and stressful looking at peoples’ faces when we introduced these collaborative discovery processes, as it was very clear the outcomes we needed to achieve but it wasn’t clear how to get there. You gotta love this feeling!
It’s important to constantly manage participants’ expectations as a way to keep engagement high, as the process often goes from well-structured to hectic and back to structured again. It’s also important to do quick recaps in the beginning of each session and quickly review at which point of the process we’re at.
The way we approached the problem
What was the best and fastest way to discover a first version of the marketplace that made sense for our customers?
First, we needed to know way more about the business, the history, the motivations, the goals, the logistics, the people (those both behind it (selling) and using it (buying)), and the customers.
Before we started, we had all sorts of different meetings, gathering as much information as possible, building a better sense of the business, competitors, references, current pain points, and needs. Each meeting was scripted and prepared ahead, so we could use everyone’s time more efficiently, extracting as much relevant information as possible. It’s beautiful to participate in meetings with strong, well-informed people from both sides, asking the right questions which trigger good answers.
Scripts and findings were all documented and made available to everyone in Confluence.
Due to time constraints and the pandemic, we weren’t able to go out and talk directly with customers. Using marketing and customer support data, we did our best to assume some of our customers profiles. We knew we could validate and clarify our assumptions later.
Thanks to this, we felt ready to dive into a collaborative process of discovery.
We chose the Design sprint as it’s a well-structured, time efficient, comprehensive method for collecting a large amount of input on a problem to help define a Minimum Viable Product (MVP).
Very quickly — a design sprint consists of 5 steps completed over 5 days. This is a concentrated and efficient way that uses the design process to solve problems addressing business and customer needs involving the necessary stakeholders.
Roughly the steps are:
- Understand / Share / Empathize with our users.
- Diverge / Sketching, creating solutions and new concepts.
- Decide which solutions better address the user needs.
- Prototype the solutions.
- Validate with users the solutions.
We’ll go through these steps in detail as we explain the way we applied it.
Scenario
We were 12 people scattered across 4 time zones and countries, all working from home due to the ongoing pandemic. It wasn’t practical for us to be connected for 7/8 hours a day as if we were in the same office for a week. We needed to make some adjustments to the standard design sprint to be efficient, promote collaboration and produce a positive outcome. Thus, we decided to break each step into 2 daily sessions of 2 hours and make any adjustments as we went.
We also decided to break our large team into smaller teams. We created 3 multidisciplinary groups of 4 people. After each break-out session, we would all come back together again to merge findings into one single consolidated view. We strongly recommend this setup as it’s easy to manage conversations and discussions in smaller groups.
First, we described all activities and the structure of our sprint ahead of time and shared it with the full team. Having so many people with different backgrounds, some of whom had never been exposed to such methods, could lead to miscommunication, generate frustration, and lower engagement.
To collaborate remotely, we selected Miro, for its ease of use and performance. It turned out to be a great choice.
Next, we structured boards to match the goals of each session and to create templates so participants would have a good and familiar starting point.
Doing these workshops fully digital demands more preparation but also allows for some simplifications as it’s easier to replicate components, copy, move, or duplicate content across different sessions and boards. But principles should remain the same: Never delete or edit any generated content.
As Day 1 was getting closer, some level of concern and discomfort was settling in. I’d done this before, but never remotely and with this level of responsibility. But that’s the fuel for growth and innovation, right?
DAY 1: Understand / Share / Empathize
Quick recap — This is where we got to know our customers and future users.
As mentioned, we weren’t able to talk directly with users beforehand but we had a good understanding of the client’s customers via the existing business.
After doing an introduction to this session mechanics and checking if all links and credentials were working properly, we started the activities.
This day was all about defining our users from both sides, buyers and merchants/sellers. We started by creating as many user profiles as possible, following a common template: name, gender, age, location, occupation and most importantly needs and goals. Although the most interesting information for us was about the needs and goals, the rest would contribute to the demographics.
We split into smaller groups, and after 30 minutes, we had a very interesting number of profiles. We then proceeded to merge the similar profiles within each smaller team, mostly based on needs and goals. Later we did the same with the results from all three teams.
Our next activity was to list and aggregate all needs from both buyers and sellers followed by creating a list of categories based in groups of needs. We ended up with a list of 24 categories for buyers (e.g., fast delivery, latest trends, product suggestions, bulk order, influencers, etc.) and a list of 10 categories for sellers (e.g., reports, inventory, support, etc.).
This was a good first day. Only four days left.
DAY 2: Diverge / Sketching
A quick recap — We got a bunch of needs and goals from users from both sides organized under categories and some user profiles (we will park that for now).
The second day was mostly about addressing our now identified user needs. We had total freedom to come up with solutions, concepts, and ideas to put under each category from the previous session.
This was a bit overwhelming at first due to the number of categories. When our participants looked at Miro that morning, they saw a “train” of 24 categories for buyers to fill with ideas. But this is where the diversity of our team kicked in. We made it very clear that we didn’t need to generate solutions for every single category if we didn’t have anything relevant to contribute. Thus, people only suggested solutions and ideas where they felt more comfortable according to their background and experience. In the end, we got a diverse and rich group of stickie notes on every category.
This was the morning. In the afternoon, we got together again to merge and consolidate our ideas, solutions, and new concepts. Participants felt more confident looking at the results, seeing the number of grouped ideas shaping what would become specific features.
Here’s an example:
Buyers category: Fast delivery (with 6 mentions)
Needs:
“Needs to buy last minute items (e.g., gifts)”
“Wants to have the same mall experience, to buy and receive “immediately” their clothes, fast delivery”
Solutions:
“Premium subscription (prime) for better shipping options ‘for free’”
“Integrate logistics partners tech to provide time slot deliveries exposing available slots to customers to book deliveries”
One potential challenge emerged: there was still a good level of skepticism among some participants as they were looking at a big list of categories with no specific (or visible) relation between them and so many solutions scattered across those categories. 🤔
On the third day, everything would fall into place.
DAY 3: Decide
Quick recap — We had a group of needs from users which generated a set of categories that were addressed with a list of solutions or new concepts. We also had a list of user profiles that needed to be focused on.
As it wasn’t viable to have a digital product focused on all possible user types, it was necessary to select a few (or one) and target the product to that group. Later, it was possible to slowly scale to a larger audience. This meant going back to the first day to look at our created users’ profiles and see which was the largest group, which had more needs and goals. From there we refined them and gave them labels (i.e., Big spenders, Influencers, Brand managers). We made a clean list of these groups of users and voted. Two user types received within buyers and sellers received the most votes.
We identified Influencers and Big spenders as our main buyer profiles, and Brand managers and Content managers profiles for sellers.
The first decision of the day was made. We had a segment of users to focus on. Now we could start creating a journey for them, where the suggested solutions could, in a structured way, solve their needs and goals.
Next: How to create user journeys with the data we already have?
From a buyer perspective, all digital multi-brand marketplaces have pretty much the same core moments of the user journey — moments like Awareness, Exploring, Finding, Ordering, Paying, and Delivery. From a seller perspective, there are moments like Onboarding, Product management, and Customer support. We built user journeys for all four profiles using the categories we already had, most of which fit like a glove to those overarching moments of the experience.
Categories like “Latest trends”, “Product suggestions”, and “Curated content” fit really well under the Explore moment.
The next part of the exercise was to select the solutions under each category better suited to that user profile and “drag” them to the user journey. As an example, for the Influencers profile, it made sense to drag the “Create my curation board section” under the “Curated content” category inside the Explore moment.
In the end we had four great user journeys that would address the needs and goals we’d identified in the beginning directly. But the day wasn’t over yet. Next came the most specific part of the activity: We needed to turn those solutions from each category, under each moment, into actions and features that will shape the user experience of the marketplace. Again, using the Influencers profile as an example, a solution like “Affiliate program for influencers” needs a feature like “user profile with sales made through his/her listings (commission)”.
This was one of the most rewarding days for our team in the design sprint. We were in the middle of the process and, after just three days, we already had a very clear idea of the features that would define the product. Everyone felt they were part of the process and had direct input and ownership.
There was another decision we had made that that we would just focus on buyers (for the sake of time). Later during development, we would come back to the seller’s experience (if this was a good or bad decision is a whole other story…).
Each day there was some off-line work. The designers spent time tidying up the boards, refining the findings through better visual arrangements, and preparing for the next sessions.
DAY 4: Prototyping
Quick recap — After 3 days we’d identified our main user profiles, “listened” to their needs, created solutions to address those needs, and designed their user journeys. Not bad!
This day was one of those that needed to be different from a face-to-face workshop like you would do in a traditional design sprint. The outcome would be exactly the same (i.e., a fully working prototype) but the mechanics were different. Due to the time available for participants to be online, their dexterity with the collaborative tool and the necessary preparation work (visual components for wireframing), we agreed that the designers would do this step off-line and then have a session where the prototypes were presented for feedback.
Earlier that day, two of us got together at the office (the only non-remote moment) and started scribbling on the white board using the findings from our previous sessions. Core pages and moments took shape. While one of us continued with that exploration, the other went looking for a marketplace wireframe template so that we could be faster creating the initial layouts. You gotta love our design community, there are good people creating great components that make our lives easier and allow us to focus on what’s important (btw, we bought the templates).
To keep the spirit of the design sprint, we did all the necessary pages in one day… what a ride! This was a Friday. On Monday, the third designer joined us, and we refined a few more of the pages and a few hours later we were presenting the flow to our fellow participants, discussing details as we went.
We exported the pages to Miro so we could connect them with our user journeys to make it clear where our pages and features came from. This was a great session, the mood was very positive, and the feedback relevant. Participants could see and feel the outcome of such an intense but rewarding process. The skepticism was now gone, and people were praising the process.
Participants could still provide more feedback off-line for the rest of the day.
DAY 5: Testing
Quick recap — We have a working prototype!
Actually, this stage did not happen over the course of a day. It took us about 6 days to go through all the scheduled user interviews. But let’s rewind a few days. Remember when we selected and voted on our user profiles? At the end of that day, we made a better description of those profiles, so that the client’s marketing team could start to reach out to users that, more or less, fit those profiles. We had a great advantage: A huge list of customers registered in the client’s loyalty program. It was fairly easy to arrange a good number of participants to interview.
The three designers split tasks so we could move faster. While two of us were refining the prototype, iterating over the other participants feedback, I was creating a script and protocol for the user interviews. We also prepared a template for notes. We wanted to validate the whole flow, features, and overall impression.
For cultural reasons we didn’t record the sessions, we just took detailed notes. The sessions went very smoothly, most users were enthusiastic, highly engaged and talkative, and happy to support a new product. The process was pretty straightforward — we shared a link with each user, and in turn they would share their screen and share their thoughts out loud as they performed the requested tasks. In the end, we had a few minutes where we openly discussed their experience. We had some hidden features that we only showed during these conversations to discover how much users valued them. These were more complex or new concepts that would not be part of an MVP, but it was already good to see how users reacted to them.
It was interesting to notice that after a dozen interviews the patterns became clear and consistent. We finished the round of interviews with 20 sessions, a good number to get consistent user data.
Instead of just creating a report (although all interviews were documented in Confluence) we went back to Miro and displayed our findings alongside the pages, using emojis to reflect the feeling of the user feedback. 🙂🙁
Later, after all interviews were done and findings digested and shared, we had one last meeting with all the design sprint participants. With no special agenda, we just wanted to wrap up and have an open-hearted conversation about the process and the outcome.
Overall, the feeling was very positive and everyone was happy and confident with what we accomplished in such a short amount of time. Some even said, “ok, now it’s only missing the UI and development”.
Not quite…
One of the great outcomes of such a process is the alignment between team members and the shared ownership. After the workshop, these people didn’t disappear to work independently in their streams, they continued to interact and work together, not only because they needed to, but also because they had this shared experience that bonded them.
With a better and validated idea of what the marketplace could become it was time to put the wheels in motion to build the product. Actually, we can’t really say this as this discovery process was already part of product development. It’s very important that people realize that this is part of the timeline of events that is product development. Design sprints or any other discovery processes should be used recurrently along development. It’s crucial for all the reasons mentioned — discovery, validation, and alignment.
Other processes can be used to discover and define what digital products can be — the design process is the basis of most of them. We found the design sprint suited and allowed for some adjustments. It’s important to understand the fundamentals of these processes to see how you can use them flexibly. We need to be smart using them so we can extract the right value out of the process.
The marketplace was launched in February 2021 and gained traction fast. Our collective creation is out there and performing. KI challengers not only helped discover the product but also contributed with great talent to create the foundations of the product and the company.
At KI group we are looking for entrepreneurs, solvers and creators who want to make a difference by building sustainable, user-, customer- and planet-driven business models & solutions in a constantly evolving world. If you’re interested in working in a fast-paced diverse environment on a variety of projects, companies, products and technologies be sure to get in touch with us — we are looking forward to meeting you!
