Our readers’ support is vital to the Guardian’s future, and the Supporter Revenue engineering team is responsible for developing digital experiences that engage our valued supporters. We live our company values by developing in the open, using best practice technologies as much as possible, and upholding our engineering culture while we work hard to ensure digital experiences remain usable, accessible and render quickly.
Back at the start of 2016, we were a small team inside a department called Digital Development, and we had a very utilitarian name: Membership and Subscriptions. Working alongside me was an agile coach, a product manager, a graphic designer, and four amazing software developers – two of which are still here but working in different areas. We all had the job of looking after two Guardian website subdomains: membership.theguardian.com and subscribe.theguardian.com. Both of these microsites have changed quite a bit: we’ve since closed down our membership scheme, subsumed newspaper and digital subscriptions into a “support proposition” – to give you an insight into some Guardian-speak.
My story starts with a project called “Zebra” – an early attempt to enable readers to support the Guardian through single payments. It used an emerging payments platform called Stripe, and that’s what inspired the name! Unfortunately it was never released to the public. Instead, the Guardian went on to add the successful £5/month “supporter” tier to its membership portfolio. That product’s main benefit was a weekly newsletter from the membership editor. The original, higher priced membership tier called “partner” offered six Guardian Live event tickets or books, and the “Patron” tier offered both of those as well as a schedule of receptions with the editor in chief.
In early 2016, shortly after a somewhat legendary offsite down in Bath, the Guardian made the decision to develop a single-payment “contribution” offering, much like what was being done over at Wikipedia. A small team was spun up, led by a product manager with both an editorial and commercial background. One of my developers initially went over to be part of that team, and it was he who came up with the team’s name which will forever be etched into Guardian lore: “Giraffe” – in homage to Project Zebra of course! This team developed a proof-of-concept and launched it in time for the Panama Papers. Over the next few months, the Giraffe team brought in an engineering manager and morphed into a “huddle” (what we call a cross-functional product development team) which included staff from design, user experience, user research, marketing, editorial and data disciplines. This huddle launched a brand new e-commerce website which leveraged the latest Stripe and PayPal payment platform software development kits, as well as a fork (a point-in-time code branch) of the open-source Snowplow analytics software which we called “Bolt”. It was an early success, mostly due to the huddle continually optimising not only the website, but also their ways of working. For example: only once Bolt’s data demonstrated that there was a high likelihood of achieving sustainable revenue from a new initiative (more about these later), did the team proceed to harden the underlying process and technology stack that facilitated it.
Meanwhile, back in the Membership and Subscriptions team (née Wildebeest team – for a brief period of time we copied the Giraffe team by naming our team after African savannah animals), we were frantically migrating 4 different subscriber bases of the digital pack, newspaper home delivery, newspaper voucher book and the Guardian Weekly – a weekly print news magazine for a global audience – away from a handful of third party operation houses, into what we called the “Touchpoint” platform which was used for membership. It comprised the Zuora billing platform leveraging the aforementioned payment platforms plus GoCardless, a Salesforce CRM, and various pieces of AWS middleware.
Just as everything eventually came online under subscribe, contribute and membership subdomains, and the marketing team launched various promotions for each product, an executive decided we needed a third huddle to simplify our offering as it was reported customers were getting a bit confused. They ironically called the team: Simple and Coherent (S&C).
S&C’s ultimate outcome, beyond creating the latest “Support the Guardian” branding, was to close down the supporter membership checkout and launch a new one for “recurring contributions” (i.e. monthly and annual contribution plans) on yet another brand new subdomain: support.theguardian.com. Single contributions (one-off or one-time payments) moved over to that domain, and Guardian patrons got their own microsite.
By October 2017, with only one place for the marketing team to direct readers who simply wanted to support the Guardian financially without expecting to receive any digital benefits in return, we made substantial incremental revenues – especially as contributions don’t incur sales taxes like memberships or digital subscriptions do.
However there was one more innovation incubated within the Giraffe huddle that arguably changed our industry: “The Epic” – our internal name for the after-article support message. The huddle discovered that the epic (which required far less prominence compared to our very first channel at scale: the banner) would yield an incredible conversion rate whenever the support message resonated in some way with the content or tone of the article above it. Its long form enabled the Giraffe huddle to define and periodically redefine “the product”: eg the importance of your support, and the impact the Guardian is making. The epic also kickstarted a flywheel effect as we realised that we could better sell our digital (app + web), newspaper and magazine subscriptions through supporter messaging rather than traditional benefit-led marketing messages. Although of course discounting still proved mightily effective, we had essentially discovered a whole new receptive audience for our marketers to engage with. A tide to rise all boats!
And then we found the booster button! By highlighting in the epic how many articles a user had read, and tailoring the emotional messaging based on that value, it busted banner blindness and allowed our marketers to encourage even more readers on a journey through to becoming a valued supporter.
But there was one more piece of magic we discovered, and without it we would have really struggled to maintain our flywheel: A/B testing of our support-led messaging. Here’s how it works: by using this technique for optimisation, we engage an evolutionary force on our messaging. We routinely collaborate with both marketing and editorial to speedily write and deploy “challenger” messages that are either relevant to the news agenda, or explain some aspect of the Guardian’s story. These are tested against regional “champion” messages (which I affectionately call the “alligators in the pond” – although maybe they should be called the “lions in the savannah”?) which contain themes that best describe why a reader in that corner of the world should support the Guardian. If a challenger message wins a site-wide test, we incorporate its themes back into the baseline message.
It’s worth highlighting that this degree of A/B testing is only possible because our single contribution checkout is far simpler compared to other products – much less personal data needs to be collected – requiring the reader to provide only an email address along with their payment details. This means that for a typical site-wide A/B test of the epic, we are able to get enough conversion data points to achieve statistical significance on the majority of those tests within one week. Without the single contribution product remaining in our portfolio, achieving significance would take nearly a month to reach.
Single contributions though come with a business problem that recurring contributions and subscriptions don’t: each cycle we needed to re-market to our previous contributors in order to drive repeat support and raise the average revenue per supporter (i.e ARPU). This challenge led to us working on probably my most cherished project over the last seven years: contribution reminders. It was my favourite because it followed the archetypical Agile software development approach: we undertook extensive user research; prototyped the solution; launched a proof of concept on the thank-you page to confirm that it was a compelling feature; swapped that out for a minimum viable product which let users choose their reminder months; agreed to manually copy those email addresses into our Braze email service provider each month so as not to over-invest in automation; tested how many reminder emails it was best to send during that month; before finally verifying that the feature achieved its business goal of increasing ARPU.
Then we rebuilt it a year later! This current incantation operates without any manual intervention, and enables you to select recurring reminders too. Since that final rebuild in early 2021, we have reminded just under one million readers via this mechanism, and more than 17% of those have contributed more than once.
In 2017 we were thankful for your support, and as we stepped up your ability to support, well beyond the inflection point in April 2018 where “Reader revenues, including income from the sale of the Guardian on newsstands, now exceed revenues generated from advertising”, we said thank you again, before entering a period which I call “The Banner Wars” …
In April 2019, the S&C huddle had been renamed: Subscriptions, and alongside the Contributions huddle (whose Giraffe name had been finally dropped), each were facilitating revenue lines that were directly competing with one another. By mid 2020, each huddle had their own distinct banner, and were vying for whose should popup first! By November we had added another section to our tooling to coordinate this banner display logic. We also helped the Apps team to link up to Braze to manage their “friction screens” and epics too!
Sitting alongside the two product-oriented huddles was the honourable Supporter Experience huddle – a vestige of that original Membership and Subscriptions team. They comprised staff from retention marketing, customer service, as well as user experience and design, had their remit covered supporting the fulfilment of our physical products and digital entitlements, the Salesforce-based contact centre screens, marketing of similar products to existing supporters, and many other customer service journeys such as onboarding, payment failure recovery and holiday stops. This huddle built what I believe to be the most advanced date picker on any website in the world to enable Guardian Weekly subscribers to manage their holiday stops in the Manage My Account area.
Today, the product separation is long-gone and the now four Product and Engineering huddles are oriented along the customer lifecycle. One half of the Contributions team split into Marketing Tools who developed additional web message channels, advanced use-cases for Braze, and metered access to the app. The other half joined the Subscriptions team to become the Conversions team who recently launched our latest Supporter Plus checkout which mixes recurring contributions with digital benefits. The Supporter Experience huddle is now called Retention. They recently built our wonderful new Help Centre, driven from Salesforce Knowledge Base articles shared with the contact centre. Last but not least, we have the Engine team who support price rises, fulfilment, compliance and data analytics use-cases.
Several huddles work together during what we call “moments’’ when we coordinate with the marketing and editorial teams to produce bespoke banners that tie in with the news agenda or amplify a particular moment in the Guardian’s history. The frequency of these led us to upgrade our frontend architecture in June 2020 to speed up their development, and in recent weeks we’ve made exciting improvements to our banner tooling to reduce the amount of hardcoding necessary, as well as to the end-to-end development process.
Here are some examples of those custom-designed banners: since starting with the Australia moment in June 2020, we then we launched our first climate promise banner in September, celebrated our 200th birthday in April 2021, added a goal ticker to the Australia moment banner in July 2021, trialled an investigations moment in September, refreshed our climate promise banner in October, and galvanised our US audience on GivingTuesday in November. We celebrated the New Year in January 2022, started covering the Australian election in April, returned to the climate crisis in September, supported both the US and Australia end-of-year moments, promised to continue covering the Ukraine war, to hold the US supreme court to account, and … take a breath … most recently, an engineer prototyped a rotating 3D Guardian logo banner in preparation for what will be our fourth climate moment later in the year!
I hope you agree that we have all undergone a remarkable journey?
