Kedro Brand Refresh

[Mackay031]

This is the brief, synthesised from the team workshop:

The need for a rebrand

Why the Rebrand?

• Kedro is evolving
• We want to increase awareness of Kedro
• We need to find our individual character and build our presence out there
• We’re moving forward, and branding will reflect that
• There are heritage QB Labs design/marketing decisions that we want to break free from
• Lack of clear Brand Guidelines leads to inconsistencies across different touchpoints
• Kedro icon is difficult to use on different mediums e.g. requires different colour backgrounds, is sometimes hard to see
• All of our assets have a different look-n-feel
• There is confusion around which brand assets to use when
• Brand assets are not extensive enough
• We need a more flexible set of assets that can work with different mediums: Website, Social, GitHub, YouTube etc.
• Heritage - mixed use of "marketing assets" and product assets
• Assets that we use to talk to users are not aligned e.g. website, documentation and deck, this leads to confusion
• Clarify our messenging
• The definition of some things in the Kedro messaging have changed over the years (e.g. "production")
• We need clarify why we're not an orchestrator?
• The value proposition isn't that clear to some users/gets misrepresented in conversations (e.g. by less technical APs)
• What is part of Kedro and what is not
• Competition
• We are competing with some very established brands e.g. MLflow, Notebooks, Databricks, Cookiecutter Data Science (Project template), Ploomber, dbt (for SQL users), Couler (from some OS user), Intake (catalog), Dataiku, MetaFlow, ZenML, TFX ecosystem (especially for DL), Hydra (sort of), Dagster + Prefect, DVC, Orchest, Airflow (sort of), Neptune.ai and Targets (for R users)

Who is struggling with the current state of the brand

• Kedro team when we have to implement the branding
• Open source contributors
• CSTs when they have to speak about Kedro
• Users looking to put the Kedro logo on a deck, and then wondering which one to use

The challenges

• Agreeing on a direction (or name?)
• Careful balance of not abandoning what has been successful up until this point
• although there is much room for improvement, as a team, we like the name ("name is not something 'AI'"), elements of the logo ("clean", "geometric" and "simple"), the yellow
• Users might not like our new branding
• Balance between doing something different but still being recognisable to users as Kedro
• Need to appear more open source and less corporate (rebranding sounds kind of like a corporate thing to do)
• Implementing the new brand in certain tools might lead to technical challenges e.g. restyling docs on Sphinx
• People might keep using the old brand, not update existing assets, etc. (e.g. PerformanceAI is still on decks)
• Virtual collaboration with brand implementation might be tricky
• The relationship with McKinsey
• it does give us marketing power to some extent, but it's not known or trusted for tech
• we don't want to alienate the McK community of users which is very valuable
• create tension with McKinsey and their stakeholders

The reward

• build trust and respect with the community
• a brand made for a ML product from the ground up
• an easily recognisable, easy to implement brand will lead to greater brand perception, greater adoption and entice new users
• less confusion around the brand assets will lead to more solid executions, building community ownership
• open new channels of marketing, again increasing adoption and enticing new users
• lead to more independence, allowing to stand alone from McKinsey
• lead to more team pride and ownership

Scope

Current and Future Needs

• Current touchpoint needs:
• Application (Kedro Viz and Experiment Tracking) https://demo.kedro.org/
• Website https://kedro.org/
• Kedro Docs https://kedro.readthedocs.io/en/stable/index.html
• Presentation decks https://mckinsey.box.com/s/9z2478akpmja7o8y9wy4bkmfhin9nrtq
• Merch (t-shirts, stickers)
• External plugins https://kedro.readthedocs.io/en/stable/extend_kedro/plugins.html?highlight=plugins
• Medium blog posts https://medium.com/quantumblack/getting-started-with-kedro-67edcc316f6a
• GitHub https://github.com/kedro-org/kedro
• Other external touchpoints e.g Slack, Linux Foundation https://www.linuxfoundation.org/projects
• Press pack

Future touchpoint needs:

• Kedro blog
• Youtube channel or training material on Cousera
• Kedro book
• Social media; specific to Twitter
• Kedro con and other conferences
• Linkedin
• Kedro Udacity/Udemy

Defining success

What success looks like:

• Competition and integration with other tools

  • Kedro stands out and picked over competitors, taking a larger share of the market
  • Kedro is the most established data science framework
  • Becoming part of the core tech stack at a well-respected tech company
  • Clearer understanding of how we fit in compared to orchestrators
  • Kedro-Viz becoming the frontpage to every company's data science code
  • Large ecosystem of thriving plugins
  • Technical Steering Committee of 10 external companies
  • Kedro adopted over Mlfow by users everytime (especially on Experiment Tracking)
  • Easy to integrate with other tools that users want to use, e.g. DVC, Neptune (even if they're party competition)

• On client engagements

  • Nobody saying "the client refuses to use Kedro"
  • Clients already using Kedro when CSTs engage
  • Community
  • Build trust and respect within the community
  • A brand made for a ML product from the ground up
  • An easily recognisable, easy to implement brand will lead to greater brand perception, greater adoption and entice new users
  • Less confusion around the brand assets will lead to more solid executions, building community ownership
  • Open new channels of marketing, again enticing news users and candidates for hiring
  • Leading to more independence, allowing us to stand alone from McKinsey
  • Excellent user support
  • Building a large community of external advocates
  • Take <5 min for a new person to understand why to use Kedro and be on board--not a pain trying to convince them
  • More advance users - should be simple to start with, but able to support advance usage (when performance is important, no need to fight with the framework)
  • Satisfying current users, gaining more unsolicited positive comments and Github stars
  • No more user questions that could be searched within 10 seconds

• Learning

  • Kedro used in academia (complex project + PhD Student with 0 software experience)
  • Kedro taught through courses/uni
  • Kedro in a book (like you would have numpy, pandas, sklearn)

• Kedro is a featured tool on all job adverts, helping with recruitment

• Helps us achieving all OKRs

• Beautiful branding that makes us all happy every day, building team pride and ownership

Project timeline

[jmnunezd]

This post is amazing! looking forward to this 💯 !!!

[tynandebold]

We've shipped the rebrand work.