Bloom Box is an online flower shop aiming to offer their customers a unique and fun experience when purchasing customisable bouquets and flower arrangements.

Based on this, as a long-term goal, Bloom Box wishes to offer their customers…

“…a shopping experience that is as fresh and fun as the flower-boxes.”

Google Venture Design Sprint Methodology


What flower buyers and experts have got to say


Customers seem to have an overall negative experience with the big online florists such 1800 and Teleflora.

Shopping locally is appealing because it’s more personable and one can see, touch and smell what the final product will be.

  • Age range: 25-34
  • Preferred method: online (41.2%)
  • Price range: $35-50
  • Most valued: beauty and quality of arrangement

Experts find it hard to meet demand peaks and manage inventory due to perishable nature of the goods.

  • Customers come back when the relationship is built on trust.
  • Men rely on advice more than women.


Some of the pains shoppers experience…


  • Impersonal customer service
  • What you see might NOT be what you get


… and some gains they’d love to have


  • Assistance when they’re unsure
  • Recommendations
  • More information: flowers and their origin, durability, allergies.


We made assumptions and we questioned them


Given what we know so far, what assumptions could we make? What questions could stem from them?



  • Our users will be able to see what the final product look like at any stage of the purchase process
  • Users will have personal support when choosing their product.
  • The online shopping experience is as intuitive and natural as it would be locally



  • Will users have the option of previsualization of their product?
  • Will users have personal support when choosing their product?
  • Is the buying process intuitive and natural?


Lesson learnt.

Initial hesitations in this exercise were dispelled when we understood that any assumptions we make, whoever we ‘impersonate’ (the business, the user) are customer-centred regardless.


Delving a little deeper – The Map


High-level steps that move our chosen actors to the final goal

We decided on an initial distinction between experienced users (they know Bloom Box and online shopping) vs newbies.

We included a 3rd person, a customer service rep, potentially filling playing the role of ‘virtual shopkeeper’ who would help out, if required.


How Might We… ?


Having mapped out the possible high-level tasks our customers need to carry out to reach their goal, we then jotted down a lot of questions that were meant to deconstruct the bigger problem and identify smaller focus areas in which we might be of assistance.

  • Completing this exercise individually allowed us to:
  • Produce a high number of questions
  • Group them by their the pain point or wish they addressed and easily identify several categories

Having two votes each, we chose our favourite HMW questions.

One category in particular, Customise & Preview, seemed to demand special attention.


How Might We – Questions and Categories


Our Sprint Target


Looking back at our map, it was clear that the target area we wished to address was the ‘Select Product → Customise’ steps.

In a quick discussion, we decided that the two should be combined in one task.

Divide or swarm? We opted for swarming: the step was ‘Customisation’ and each one of us would go wild on it.


Lightning Demos – time to find inspiration


Each one of us had to find two or three companies, websites or designs that contained an interesting ideas and ways of handling customisation.

3 minutes was the time for each demo.

An appointed note-taker then sketched each idea on a post-it.

Once we had all the sketches on the board, we took another good look and tried to capture the main ideas next to each drawing.


Sketching time!


Aptly inspired and with ideas swimming in our heads, we were ready to start focusing on the next big phase, sketch, itself made of 4 stages:


Look back at your notes so far (time – 20 minutes):

      1. The goal
      2. The interviews and survey
      3. The map
      4. The demos


Jot down ideas (time – 20 minutes):

      1. in words and little sketches
      2. a mind map
      3. doodles


The Crazy 8s (8 sketches, 1 per minute. Time – 8 minutes)

      1. look back at mind map and doodles
      2. think of a feature, a flow, a story, capture the core ideas




    1. the 3-panel self-explanatory storyboard (time – 30 minutes)

Mind Maps and Crazy 8s


Sketch – The 3-panel Storyboards


Having gone through a phase of exploring ideas, first without constraints, then with the ‘crazy’ constraint of one minute, we were ready to put some method into our madness.

Tell a short story with a bit more detail, but still centred on the core functionality

Must be self-explanatory


Time to Make Decisions


The mantra for the day was: Decide today, build tomorrow

The idea is to explore all 3-panel storyboards silently on an individual basis and refrain from judging the designs just yet.

  1. The Art Museum. Like in a gallery, we analysed all designs and asked clarifying questions on post-its placed below each board.
  2. The Heat Map. Next, we had a good supply of sticky red dots which we placed next to our favourite designs, features or ideas.
  3. Speed Critique. The Decider looks at each design and names its core ideas. These are placed on post-its near the design. Everyone discusses and interprets the board and finally the creator is allowed to intervene and clarify.
  4. Straw Poll. 2 dots each for a final vote.
  5. Supervote. In the event of a tie, the Decider makes the final decision on the design to implement

Art Galleries, Heat Maps, Ideas and Dots.


A few lessons learnt


It’s hard not to criticise from the start. We had to remind ourselves that the first moments of observation were not about judging

It’s hard for the creator to keep quiet while the others discuss your design

Avoid ending up in ‘yes-no’ fights, they’re dead ends

It’s in the nature of the exercise to raise criticism, nobody should take it personally


The Storyboard – Imagine a User in Action


Having decided on a design to adopt, it was now time to develop it in more detail, visualising it in action in the real world.

The main idea was to imagine a scenario in which the user needed to use Bloom Box’s online services and follow her through the process, from the moment that triggered that need, until the finished product was ready to be purchased.

Worked well

  • We stuck to and developed the 3-panel design ideas, referring back to it every so often
  • We went back to the long-term goal and sprint questions to ensure we hadn’t departed from them
  • Fairly good time keeping, we were able to get ‘unstuck’ and move on when we felt discussions were getting too detailed
  • We showed flexibility and willingness to give in if some of our ideas didn’t fly

Could’ve done better

  • Getting stuck. It happened frequently, we should have tried to avoid it more



Time to Prototype – The RULES of Prototyping


Cutting-edge technology was utilised in the creation of our state-of-the-art prototype solution: paper, scissors, markers, sharpies, post-its and some sellotape.

We only had one version, but if two or three variations are built for some A/B testing, then assigning each solution a name is might help to tell the solutions apart.

  • Faking it – add detail until your creation is as realistic as possible
  • Time limits – use time constraints as a motivator to get unstuck and move on until completion
  • Anything can be prototyped
  • Disposable – don’t get attached to this baby as, if the user doesn’t like it, you’ll throw it out and start again


Time to Prototype – The ROLES in Prototyping


DIVIDE ET IMPERA (lat. divide and conquer)

Depending on the size of your team:

  • Makers – 1 or more. These are the ones drawing, cutting, colouring, etc.
  • Stitcher – 1. Makes sure the makers have the assets they need, pays attention to every detail for maximum realism.
  • Writer – 1. Handles the text parts in the prototype.
  • Asset Collector – 1. Finds resources (internet, real world) for the makers. The Stitcher collects the assets from the AC and passes them on to the Maker.
  • Interviewer  – 1. Prepares for the interview with the usability testers, sets the scene, gives the user tasks and asks questions.

We decided to wear at least two hats, e.g., Maker and Stitcher or Asset Collector and Maker, to handle workload and time left.


The Prototype – Things we did well


Collaboration. The only roles that were somewhat removed from the actual making of the prototype were the Writer and even more the Interviewer. However, the Writer would keep an eye on progress and satisfy the request of the maker/s and the Interviewer jumped briefly in at the end and adjusted the script based on final results.

Getting stuck and unstuck, using the same tactics as during the rest of the week, i.e., acknowledging we had reached a dead end and it was time to move on. Foer instance:

Crazy 2s or 3s

Referring back to the original long-term goal and its key principles (easy, fun, personable)

Completing by the deadline with an interface that actually looked simple, clean and easy to manoeuvre


The Prototype – Things we could do better


Clearer decisions at previous stage. When creating the storyboard, we should have expanded a little more on certain parts of the functionality we decided to leave our of the story, e.g., personalised support.

As we left it until the last minute, our third principle of ‘personable’ wasn’t quite fleshed out properly and we’d need to iterate on it

Better communication before starting to ‘make’ to avoid misunderstanding later on and some frustration stemming from it. E.g., we didn’t quite agree on the mechanics of some step such as how to select flower types and flowers.


the prototype

Doing a dry run of the interview was also very valuable to have an initial feel for the interface and to catch some very last minute details we missed out.


Testing with Real Users


The moment of truth finally arrived on Day 5, time to sit some users in front of our state-of-the-art technology paper-based interface and get some open and honest feedback.

We could run four 15/20 min long recorded interviews which we thought went quite well overall and gave us an indication that our solution was simple, fun and somewhat fresh.

A quick debrief after each allowed us to collect our feedback (negative, positive and neutral) on post-its that later ended up on a feedback board

Generally not departing from our guiding principles – except for ‘personable’ – had been our main concern and the usability tests confirmed we hadn’t generally deviated.


Debriefing, Collecting and Analysing Feedback


To quickly visualise what type of feedback we managed to get, we put our collective post-its on a white board.

We then proceeded to categorise by type of feedback:

  • a red dot means ‘negative’ (features to iterate on)
  • green, ‘positive’ (a positive comment, something that worked well)
  • no dot, ‘neutral’ (something to keep in mind for now)

Following an initial categorisation which confirmed our initial impressions of not having strayed from our main goal, we grouped the post-its by their nature and identified which part of the functionality they referred to.

We then proceeded to identify our next steps to improve the interface where it needs more work.

white board

Need iterations


We’ll need to reiterate on:

  • Text – too small, need for clearer and shorter instructions
  • Countdown (or up) for number of flowers purchased so far
  • Personal Support functionality (chat, help) went unnoticed
  • The slider to select the size of the box might have to be reconsidered.


Next steps


Next – we agreed we might have to go back to some HMW questions and some lightning demos for functionality that was misunderstood.

Some crazy 8s and A/B testing should also be done until the above points are tackled and ready to be tested again for usability.



Thanks to my stakeholders team in this Design Sprint

Sandro, Verena, Matilde, Laia and me.