Design team leads

Regardless of size, each design team benefits from a single point of authority and leadership, an individual with vision and high standards who can get the most out of the team. This is the most important role on the team, and the hardest job to do well.

Team leads must be able to:

Manage down

Design leads are responsible for overall team performance. They need to create a space (whether physical or conceptual) where great design work can happen. They must coach, guide, mentor, and prod. They address collaboration challenges, personality conflicts, and clear mandates, and peoples emotions.

Manage across

Design leads coordinate with product leads, business leads, technology leads, and people in other functions in order to make sure that their teams’ work is appropriately integrated with the larger hole. There must also be able to credibly pushback on unreasonable requirements, and goad when others claim that the design teams work is too difficult to be delivered.

Manage up

it’s crucial that these leads are comfortable talking to executives, whether it’s to explain the rationale behind design decisions or to make the case for spending money, whether on people or facilities. Design leads must present clear arguments, delivered without frustration, that demonstrate how their work ties into the larger goals and objectives of the business.

In short, the best team leads are crimination of coach, diplomat, and salesman. [..] They oversee the end to end experience, ensuring that user needs understood, business objectives are clear, design solutions are appropriate, and the final quality is high. To achieve coherence, they must integrate efforts across product design, communication design, user experience research, and content strategy. They are responsible for articulating the design vision shared not just by their immediate team, but they cross functional partners as well.

P Merholz & K Skinner: Org Design for Design Orgs, 2016, p.52-3

Centralized Partnership: The Best of Both Worlds

Being centralised is too slow and disempowering. Designers enjoy the variety and collegiality, but ultimately feel stifled by the lack of impact. Going decentralised initially improves matters, in terms of both speed of release and quality, but all the time designers grow restless and feel lost, and customers struggle with a lack of cohesion. Teams are thus tempted to re-centralise, but fear they’ll just end up in the same place.

It’s like the Kobayashi Maru*, but there is a lesson to take from Captain Kirk. There is a third way to run design organisations, one that is slowly gaining wider acceptance. There is no standard name for it, so we will collet that Centralised Partnership.

For most companies, it makes sense to have business and product teams decentralised. There are very real benefits of autonomy, primarily speed of execution, and decision-making let’s takes into account the actual on the ground, day-to-day reality. It’s a force that mostly does good. But it can lead to entropy, where the sum of the seemingly right decision is made in isolation leads to a fractured use experience. A centralised design organisation serves as a beneficially opposing force whose holistic perspective can ensure coherence. It’s not uncommon for multiple product teams to work on related adjacent or even identical stuff without knowing it. In the centralised partnership, the ongoing communication across the design organisation makes this a parent early on and those teams can figure out together how best to tackle it without wasting effort.

The centralised partnership attempts to realise the best of both the decentralised and centralised models. From an organisation chart perspective it is centralised with all designers in one organisation reporting up to a single point of leadership.


Unlike the decentralised model, which orients on staffing design at the individual level (as many decentralised teams have only a single designer), the centralised partnership takes a team mindset from the outset. Just like flocks of gees are safer (thanks for the many eyes) and expend less energy (thanks to the reduced drag) then geese flying individually, a gestalt occurs where a strong design team can accomplish much more than the same number of designers working on their own. A centralised partnership design organisation is composed of these sub teams, and the bigger the organisation, the more teams there are.

*Kobayashi Maru is the Starfleet Academy battle simulation that tests how officers handle a situation where both choices lead to loss. Captain Kirk, the stud he is, hacks the test, creating a third way that leads to success.

P Merholz & K Skinner: Org Design for Design Orgs, 2016, p.49-51

High contrast

However, the contrast can be too high in some cases. People with dyslexia have shown to be able to read faster when color pairs have lower contrasts, plus high contrast is known to hurt people with photosensitivity & chronic pain (I can attest to that myself as someone who is suffering from a combination of photosensitivity and migraines).

Twitter’s new font and Last of Us 2: an accessibility lesson to be learned

The 5 most important questions to ask when starting with a new dataset:

The 5 most important questions to ask when starting with a new dataset:

#1 – Purpose
What was the purpose for which this dataset was collected?

#2 – Competence
How competent were the people who collected this data?

#3 – Agenda
Is there any reason to suspect that the agenda of the people who collected it might bias the data?

#4 – Clarity
How clear is the documentation accompanying the data? Can you be sure you know what happened in the real world when the data were captured?

#5 – Processing
Are you looking at raw data or has the information been transformed in some way? Does this render the data unsuitable for your needs?

Cassie Kozyrkov, Chief Decision Scientist at Google, Inc.

New Usability Heursitics

<Google translation from Danish>

9 overall principles of usability can help ensure that users can complete their errands on websites in an effortless and satisfactory manner. The 9 principles are:

  • Simplicity
  • Consistency
  • Detectability
  • Structure
  • Clarity
  • Control
  • Usage signaling
  • Tolerance
  • Ergonomics


As simple as possible. Not overwhelming. No superfluous action steps.

Only what is needed
Redundant elements make a user interface unnecessarily cumbersome to overlook and increase the risk of errors. Items that do not support a mission or purpose should therefore be eliminated (not just hidden).

Limited number of options
Many choices mean that it takes longer to choose between them ( Hick’s law ), and sometimes it completely prevents users from making a choice ( Paradox of choice ).

Easy interaction
Interaction with unnecessary action steps increases time consumption and the risk of errors.

Default settings
Good default settings ( defaults ) simplify interaction.


Uniform and recognizable presentation and interaction. Corresponds to users’ expectations.

User expectations
The user interface and system responses must be consistent with users’ expectations. Users have a mental model of how a website works, even before they get into it. Mental models are shaped i.a. of users’ experiences with other websites.

Internal consistency
Uniform elements and features should be presented and work in a consistent manner on the same website.

External consistency
The user interface should be designed in accordance with conventions. It makes it easier to learn how to use a website for new users.

Consistency over time
Drastic redesign makes use difficult for returning users, while ongoing minor changes better ensure consistency over time.


Visible elements, not hidden. Designed and positioned so that they are easy to spot.

Graphical user interfaces ( GUI ) are based on visibility. If items are hidden, users should either look for them or remember their location. It is easier to recognize elements than to remember them.

The location of items can make them difficult to spot. This may be because an element is located on the periphery of users ‘field of view – or another element that acts as an equally good bid attracts users’ attention. Although items are not hidden, they can be difficult to detect.

Size, clarity and highlighting affect whether users notice an item.

Scroll-supporting design
Large images, boxes, horizontal lines and white air can give users an experience of having reached the bottom of a web page ( illusion of completeness ) and thus stop them from scrolling on to elements further down the page.


Division, categorization and grouping according to users’ needs and logic. Natural order.

The feeling of overwhelm can be minimized by splitting when simplicity cannot be achieved by removing elements.

Lack of categorization, as well as a categorization that seems illogical, complicates navigation.

Items that belong together should be grouped.

Logical ordering of items in an overview makes it faster for users to find what they are looking for.

Meta data
Taxonomies, keywords, and other forms of meta-data can help users locate large amounts of information when searching.


Understandable, accurate and unambiguous communication. Clearly several elements. Clear context.

Text, links, labels and menu items create confusion if they are incomprehensible, inaccurate, ambiguous or too general.

The difference between several similar elements can be difficult for users to understand, even if words and descriptions in isolation are clear. Comparison and selection presuppose that the difference in elements is clearly described.

Information available
Incomplete or missing information – including missing help texts and missing feedback – gives ambiguity.

Context can affect understanding and clarity. The combination of texts, images, menu items, etc. may result in a different decoding than if the elements were presented individually. The same concept can have different meanings for different user groups.

Website navigation, texts, images, etc. must together make it clear what a website or a web page is about.

Visual clarity
Visual language can also be clear or obscure. Icons, images and highlights can be more or less self-explanatory. (Clarity can also relate to other senses).


The user and not the system controls. Appropriate pace. Actions can be undone.

User management
The experience of having control is the prerequisite for users to feel that it is nice to use a system. Automatically changing elements can be distracting.

Appropriate pace
The pace of interaction must be controllable by users. There is not one pace that suits all users.

Undo options should be available when users make mistakes or change their minds.

Signal for usage signal

Usage signaling
Instructions for use. Interactive elements signal how to use them.

Interactive elements should be as self-explanatory and intuitive to use as possible. It should not be necessary to read instructions to see how basic functions are used.

Signaling of interactivity
Interactive elements should be designed so that they clearly signal how they are used. Are they clickable, can they slide , can you enter anything?

No misleading signaling of interactivity
Elements that are not interactive should not signal that they are. Users must be able to clearly distinguish interactive elements from non-interactive ones.


Sensible interpretation of users’ entries, clicks and presses. Small mistakes are ignored.

The system allows variations in user input and interprets what users are trying to achieve in a sensible way.

Error handling
Insignificant errors by users are ignored by the system.

Extended clickable areas
Click and click tolerance can be achieved by extending the clickable area of ​​an element.


Physically easy to use. Challenges and does not exceed motor and sensory abilities.

Effortless interaction
Interaction should be adapted to human physics. A user interface that forces users to physically awkward and cumbersome actions increases the risk of errors and negatively affects users’ pace and satisfaction.

Items should be large enough to be easy to hit.

Elements to be used in extension of each other should stand close to each other so that users do not have to make unnecessarily large movements.

Visual ergonomics
Ergonomics can also be about visual conditions: good contrasts and sufficiently large size of text and other elements.


/// Background

The 9 principles for webusability have been identified by analyzing 299 usability problems on Danish websites. The usability tests include websites from both the public and private sectors, intranets, mobile websites and prototypes and were conducted between 2012 and 2015.

Heuristics have been used in the design of user interfaces since the 1990s. Older heuristics sets are still inspiring to read, but they are designed based on usability issues in software. Phenomena such as feedback and error handling therefore take up a lot of space.

On websites, navigation, content and visual design play a bigger role than in software. The 9 principles of webusability make it easier to categorize issues related to these matters.

Feedback, error messages and help texts are described in many heuristic sets with independent principles. In our system, these components are just objects for evaluation just like a website’s navigation, content and functions. An error message can, for example, be visible, clear, simple, etcusa