In Designing experiences like a movie, I suggested to add two dimensions to the traditional evaluation map of desirability, feasibility and viability. The new dimensions are brand and identity, and sustainability and social impact.
In this hypothetical case study, I will demonstrate how to assess the “goodness of fit” (a term borrowed from regression) of a design concept to a company’s brand strategy.
Design Concept
I will use one of the concepts from the cafe case study as example in which I redesigned the app’s interface to look like display fridge — the first touch point that lifts up persona Jenny’s mood and brings her out of morning sluggishness. The objective is to encourage users (customers) to use the app for ordering, thus allowing cafe staff to plan customer orders efficiently. Currently, cafe staff are reacting to customer requests and cues (oven timer).
Cafe’s Brand Strategy
Assuming that our client (a cafe chain) has this brand model:
Note: This example is based on publicly available information with minor adaptation and inference. It is a concise brand model that summarises the cafe chain’s brand strategy in four statements. I have omitted details like personality to keep it simple. Please refer to this link for explanation on how to develop brand models.
Note: This case study is a follow-up to the semi-fictitious case for my course project that incorporates data, insights and scenarios from multiple brands. The Starbucks outlet I visited was intended as proxy only.
Methodology
Evaluation is both art and science.
The methodology I propose here has two main components: qualitative and quantitative assessments. Their order is not fixed and they build on each other. In some cases, we may need to do them simultaneously or interchangeably.
Qualitative Assessment
Experience, intuition, and thorough understanding of our client’s brand, business and market constitute the main “data” for qualitative assessment.
I start with assessing the fit between the target consumers of the brand (brand model) and the needs of persona Jenny whom I designed the service for. I see a good fit.
Brand Model: People going about their busy schedules need speed and convenience, but without compromising on quality, freshness and taste. A bit of indulgence will spice things up and make their hectic lives more bearable.
I then distill the design concept into its distinct characteristics and assess its fit with brand strategy. When filling in the table, we should ask ourselves about now and future. For example:
- How is the fit between the design concept and brand positioning?
- How will the design concept facilitate the brand achieving its promise?
- Which design concept best lives out the brand experience as defined by the brand strategy?
This template provides a useful tool for aligning stakeholders and explaining the rationale behind our decisions.
Unlike what market researchers (where I started my career) may suggest, I do not think we should base our assessment on users’ (consumers) feedbacks only because they do not have the level of understanding (e.g., brand strategy) needed to do this assessment. We are the best consultants.
Quantitative Assessment
Although experience, intuition and thorough understanding of our client’s brand, business and market are equally important in quantitative assessment, I will hold them back and let the data speak for themselves first.
I suggest a combination of big data and ad hoc research.
Examples of Big Data: sales/retail data, traffic data (inside cafe and in the vicinity), mobile app usage, google analytics and website analytics, facebook/Twitter/Instagram data, eye-tracking/neuroscience data, weather data, travellers data, stock prices, market caps and macro-economic data.
Examples of Ad hoc Research: concept tests, usability tests, customer satisfaction surveys, mystery shopping studies, brand health trackers, brand values and buzz monitoring.
Below are a few thoughts on how to leverage these data:
- Sync different sources of data (big data + ad hoc research data). This will provide us a foundation for analytics and predictive modelling (examples of analysis in brackets):
- Look for patterns and relationships (e.g., visualisation, correspondent mapping, correlation/regression)
- Training and testing data to build predictive models (algorithms — e.g., decision trees, classification, time series)
2. New data (from big data or ad hoc research) can be used as inputs for predictive models (algorithms) to predict on metrics such as uplift on brand health, brand values and sales, etc.
Syncing data requires specific skillsets such as SQL (relational data management). I will leave it to other experts to elaborate on this topic.
The following is a sample predictive modelling dashboard on Microsoft Azure.
An Insights Programme for Brand, Design and Business Management
My proposed methodology provides a roadmap for leveraging the trove of data a company generates and collects every day, and developing an insights programme that draws from both qualitative (art) and quantitative analyses (science). Its key selling point is linking design to brand strategy, and to financial results (sales, stock prices and market cap).
In a real project, methodologies will be tweaked as we progress. Just like in design, analytics and predictive modelling are not a linear process. The above proposal is a starting point only and will evolve continuously.
