Here’s what the blog team has been reading this week…

The cost center trap

Our first recommend is by Mary Poppendieck, author of a number of books on lean software development.

Poppendieck’s blog, The Lean Mindset, this week covered IT and the tendency of many businesses to judge the department’s efforts on cost management rather than improving business performance. This can have a negative impact on attempts at Agile working and digital transformation initiatives.

Key paragraph:

“Capitalization of development creates a hidden bias toward large projects over incremental delivery, making it difficult to look favorably upon agile practices. Hopefully we don’t have to wait for another generation of accountants to retire before delivering software rapidly, in small increments, is considered a good thing.”

Read it here


How brand storytelling began

Nick Asbury writing in Creative Review brings us an excellent history of brand storytelling, complete with some rather wry parodies of typical brand stories.

Key paragraph:

“ can see the appeal of storytelling as an analogy, or set of analogies, for what branders and advertisers do. There have been other analogies in the past – there was a long-standing military theme (tactics, campaigns, target audiences), then the built environment (brand architecture, brand pillars, brand pyramids), and a blue-collar engineering theme (boilerplates, toolkits, roll-outs). But storytelling offers an especially fertile field of analogies for what branding is about – narratives, story arcs, characters, motivations, resolutions.”

Read it here

Ritson on Facebook targeting

Mark Ritson this week looks at the impact of Facebook on the 2016 Presidential election, but with an unlikely glint in his eye. His point is that amidst all the stories of fake news and bad actors, Facebook’s sales team (as revealed by Buzzfeed) approached political parties with a segmentation of the American population (see image below) that serves as a lesson to all marketers.

Key paragraph:

“Crucially, the Facebook team started with.. attitude clusters, then moved onto behavioural indexing and then finally looked for demographic distinctions from the general market mean, rather than the more inane method of segmenting by age and gender and then making up a bunch of stereotypical bullshit about what this ‘segment’ probably wanted.”

Read it here

political segmentation by facebook

Ecommerce Best Practice Guide

Steffan Aquarone, Econsultancy’s consultant head of best practice reports, has produced a chunky and practical guide to ecommerce, covering everything from tech stacks and ecommerce platforms, to team structures and site design, by way of BAU and innovation.

Key paragraph:

“..All manner of considerations that go into customer acquisition and retention are interwoven into ecommerce, and if ecommerce is to be treated as a department within a business, then there are myriad other functions that need to be in sync with it. This guide will mention these areas too, whether it’s about the need for consistency of an organisation’s offering between online and in-store retail, or the inescapable interdependence between ecommerce and order fulfilment capabilities (sometimes leading to ecommerce managers having direct oversight of logistics and warehousing too). But in order to have a clear focus, a line must be drawn, and from when people arrive to when they’ve completed a purchase is where we’ve drawn it.”

Subscribers can download it here

ecommerce best practice guide

Solving the problem of AI bias in “black box” algorithms

Next, a piece from MIT Tech Review on the problem of AI bias in ‘black box algorithms’.

When human biases cn be baked into algorithms, how do we guard against unlawful or immoral decision-making? Well, a paper from Sarah Tan (then of Microsoft) outlines how researchers can mitigate the problem by mimicking such algorithms.

These mimics are then trained on real-world data to find out what variables were most influential in the original data set. There were some surprising results, including a loan company whose algorithm didn’t seem to place much weight on applicant income or loan purpose, despite their high risk factors.

Key paragraph:

“Critics may point out that these aren’t exact replicas—out of necessity, the researchers were making a lot of educated guesses. But if the company behind an algorithm isn’t willing to release information on how its system works, approximation models like the ones from this research are a reasonable way to get insight, says Brendan O’Connor, an assistant professor at the University of Massachusetts, Amherst.”

Read it here

The role of copywriters in a GDPR-ready world

One from our own blog now, a simple but elegant article from Daniel Saunders at Sticky Content, who lays out some questions copywriters should ask themselves in light of the GDPR. Everything from exercising caution with personal pronouns (‘we’) to aiming for a new level of transparency.

Key paragraph:

“Disclosing what information you have on someone and how you use it is an opportunity to explain the benefits. Does knowing about your customer make their experience better? Does it save them time or money? If so, say it.”

Read it here