The media industry is trying its best to adopt AI, with 80% of companies intending to invest in AI in the immediate term.
But how do you implement AI into your digital publishing company effectively, avoiding risks and huge investments? Reportedly the biggest blocker to adopting AI is lack of strategy, with 57% calling it out according to a Financial Times publishers survey.
In this blog, we look at why AI strategy is such a sticking point, and how you can kickstart your strategy with Bridged Media’s simple 3-step process. (Plus, a tip on how to begin your AI journey for less than £250.)
The AI strategy trap
The complex issues facing the media industries are actually quite simple: lack of money, lack of data and lack of skills. For instance, estimates suggest that 90% of publishers can’t afford to build an AI prototype themselves. And even if they did have the funds to build it, they lack real-world data to train and train the prototype’s algorithms.
Then there’s the fact that, aside from international mega-publications backed by serious financial heft, most media companies lack any human resources specialised in AI tech and product management.
Beyond that, the AI world is diverse, decentralised and, frankly, confusing. There is no real standardisation of AI use cases and it’s often exceptionally difficult to predict and calculate a tangible ROI.
That’s why we wrote—somewhat provocatively—in New Digital Age that “media businesses don’t need an AI strategy, they need the right tools”. Clearly we don’t mean you should ignore strategy altogether or sleepwalk into risk, but instead think about doing business better using simple and tangible tools.
So, you want to test AI in your media company, but don’t know where to start? Our AI experts at Bridged Media tend to approach these discussions in three stages.
Identify your use case
Before jumping into the AI solutions out there, first crystallise the problem you want to resolve. The chances are there’s a particular bottleneck or use case to address, whether that’s lacklustre performance on reader revenue, poor user experience due to lack of content personalisation or low newsletter subscriptions.
Once you’ve identified a use case to test, next define your KPIs. Here, you’ll probably find yourself in familiar territory, with media organisations we work with typically tracking engagement (such as bounce rates, session time and so on), as well as conversion (click-through rate, emails collected, etc) and churn or first-party data collection.
Here, we can help you learn from others’ examples to solve the use case you’re targeting, for example:
- Enhancing personalisation through recommendation engines which predict the next best action for a reader
- Processing data through AI-powered approaches like propensity modelling, visitor segmentation and insight generation
- Discovering patterns, especially those which require trawling large amounts of data and may be difficult for humans to conduct, as well as generating backlinks and automating social media
- Operating editorially, which could include anything from content generation through to simpler tasks like editing and tag generation
For example, UK-based news website The Conversation used Bridged media to build a proprietary personalisation engine and conduct AI recirculation, achieving a 15 times increase in recirculation and a 35% increase in average sessions time per user.
Completing step 1 will literally pay-off later on as organisations with clear goals see better financial returns and efficiency savings. The task also encourages some really healthy introspection across your company and should help you understand where AI can power your business model.
Decide whether to build or buy
There are two main ways to embed AI into your business: you can build it or buy it. I’m guessing you may already have a knee-jerk response to the right path for your business, but there are some useful factors to consider.
Think about your in-house resources, and ask if you already have or can recruit the skilled data scientists, engineers and domain experts needed to develop AI capabilities.
For some larger businesses this is achievable, like the Financial Times which built propensity models to double its subscribers, alongside acquiring AI-led companies to double its digital advertising business. But for small and medium-sized media companies this is unlikely, and for others it’s simply not the right fit: Condé Nast’s personalised content, prediction engine and collected first-party data is all from external vendors.
When weighing-up build versus buy, also consider your time-to-market. How urgent is it for you to deploy any AI capabilities? Building is bespoke and tends to take longer, whereas buying from a vendor is faster—just make sure you choose companies that give a quick way to pilot and test.
Pilot, measure and reiterate
So, you’ve identified your use case and chosen whether to buy or build. Now it’s time to pilot. Here’s how you can ensure your pilot is successful:
- Define your goal—most digital publishers use KPIs for engagement (e.g. time spent, bounce rate, etc), personalisation (e.g. retention rate, percentage of sessions personalised, etc) and conversions (e.g. conversion rate, emails collected, etc)
- Train your users—think about what training is needed for your team. Do you need external help from a vendor?
- Implement—create some clear product management for your pilot. Ask yourself: who is accountable, consulted and informed throughout the implementation? Create milestones and track them in one place
- Baseline—before launching the pilot, establish baseline values for the selected KPIs using historical data. This provides a reference point for evaluating the pilot’s impact
- Track in real-time—use analytic tools to monitor KPIs in real time during the pilot
- Analyse AI against human—called comparative analysis, this means assessing the performance of AI-enabled content against non-AI content to gauge the effectiveness of the pilot
- Repeat—reiterate your approach to improve your results, and continue to test and learn
Ready to kickstart your AI strategy?
At Bridged Media our mission is to empower media companies to adopt AI in an easy and tangible way. Our no-code AI solutions can help you engage your audiences and monetise them. And did we mention adopting AI ‘on a shoestring’? You can try one solution from £249—that’s a no-code, box-fresh tool to use for 3 months on your website. See for yourself.