At some point in all of our lives, we’ve undergone a change of some kind.
Whether it be getting those braces off as an awkward teenager or dramatically chopping all your hair off after a particularly heartbreaking split, everyone loves a good makeover as it signals a new beginning and a fresh start.
And while changing your look is something most of us have done, just because the look and feel of your brand may not have changed since its inception doesn’t mean that at some point, it too couldn’t do with a makeover.
We recently helped a major international brand makeover their brand and they didn’t even have to endure a back wax or harrowing eyebrow sculpting in the process. This brand was hoping for something new; a refresh if you will.
And as part of this repositioning, they were looking to change-up photography, colours, fonts and layouts. Not a minor exercise when you’ve got literally hundreds of thousands of images and web pages online.
Rather than reviewing each of their 500,000+ photos and digital assets one at a time, we generated reports out of their current content management system and then processed their visual elements to identify brand colours within a tolerance range.
With this data in hand, we could size up and figure out how to go about changing those colours to reflect their new look. We filtered images above a certain percentage and while there were some false positives, the program produced around 90 per cent accuracy.
If you’re wondering how long it takes to run this kind of program, it’s usually between a day and a week for a full download and analysis - it really comes down to the number of assets your brand currently has.
Keep in mind, analysis at this scale takes way less time than manually assessing each asset — even at two seconds each, it would take one very bored person almost 8 weeks to report back.
And ten years ago, writing the code for this kind of tool would have taken a small army of developers but as technology is forever marching forward with new advancements every day, things that used to take much longer (and at a much higher cost) can now be done in ways that work smarter and faster. We now have the capability to analyse BIG data which we can then report back to clients by informing them exactly where, and how many, images or assets they’ll need to replace in order to fully makeover their brand.
Better analysis and insights allowed for better planning and ultimately faster time to market.
The client can then manage their budget and timeline to rectify it as we’re really turning this tool into something that can report or catalogue this information — remember, we work smarter, not harder. Although our program still requires human elements to recreate the digital assets (read: the new brand colours) in order to match against those already on the site, they can easily be swapped out automatically for new images with a script and a map or dictionary.
The tool can also determine the colour theme or palette swatches of a set of brand assets to ensure they’re in theme with the new look as it has the capacity to colour audit, categorise images into colour themes, check all images with the same colours and tie in colours with objects as it can also detect objects — pretty cool, right?
However, if you only have two dozen images and not a lot of images with brand coloured offices or event banners, this type of analysis may not be for your business.
But, the tool doesn’t have to just recognise colours.
If you’re looking to identify objects, logos or even identifying faces in images, it may become more applicable for a different need.
Because of this program, and with our help, major brand reworks can be done efficiently and accurately because we like to work smarter for our clients.
Titbits and takeaways
Preparation is key - understand the beast. By having effective reporting methods set in place to deal with big data, you’ll be able to better formulate a plan of attack.
Consistency. Rebranding is a huge monetary and labour investment. Ensure you start off on the right foot and remain consistent with imagery and language from the get go.
Beyond colour. Scanning for objects, gender and ethnic diversity in photos is also possible along with extracting metadata and dimensions.