Treat practice like the Super Bowl: what Tom Brady can teach marketing teams about getting better
What we learnt at UNBOUND 2026 about practice, feedback, failure, AI and building a marketing team that improves on purpose.
UNBOUND 2026 lessons series | UNBOUND brings together ideas from marketing, leadership, customer experience and business. In this series, we are sharing the sessions that stayed with us, then asking the only question that really matters once the conference lanyard is back in the drawer: what can we actually use?
Jump to: Practise before the pressure | Make improvement specific | Use failure properly | The AI learning curve | The marketing practice loop
Marketing teams are expected to perform in public all the time. Launch the campaign. Present the strategy. Write the pitch. Build the AI workflow. Hit send. Then everyone looks at the numbers and decides whether the work was good.
What teams are given far less of is somewhere to practise before the result really matters.
That was the part of Tom Brady’s closing UNBOUND 2026 conversation with HubSpot co-founder Brian Halligan that stayed with us. HubSpot called the session “What Winning at Scale Actually Requires”, and although there were obviously plenty of football stories, the useful bit for marketers was not the seven Super Bowl titles. It was the way Brady described getting better.
Brady was the 199th pick in the 2000 NFL Draft, after years of having to work his way up the depth chart. His point at UNBOUND was that not being the obvious physical standout forced him to build other advantages: preparation, competitiveness, resilience, specific goals and an ability to learn from the reps he was given.
Most marketing teams are trying to perform before they have practised
Brady described treating practice with the seriousness of a game. Later, when the stakes became enormous, the pressure was not completely unfamiliar because he had spent years creating urgency before anyone was watching.
Marketing has a version of this problem. Someone is told to start making video, build a new nurture sequence, run a discovery call, present a pitch, use a new AI tool or write for a channel they have never used before. Their first meaningful attempt is often also the first attempt that gets judged.
If it goes badly, the conclusion can become “video does not work for us”, “AI writes rubbish”, “email is dead” or “they are not very good at this”. Sometimes the channel genuinely is wrong. Sometimes the person simply had no practice field.
The practical marketing version is simple:
- Rehearse the pitch before the prospect is on the call.
- Role-play the objection before a salesperson has to answer it live.
- Test the message on a smaller audience before putting the whole budget behind it.
- Run an AI workflow on low-risk work before making it part of a client or customer process.
- Review a piece of content with the same scrutiny you would use if it were going in front of your most valuable prospect.
The point is not to create endless rehearsals and never ship anything. It is to make sure launch day is not also lesson one.
“Get better” is not a useful goal
One detail from Brady’s talk was particularly practical. He used to go onto the practice field with a small list of things he wanted to improve that day. Not “become a better quarterback”. Specific things: footwork, follow-through, movement, technique. The ambition might have been enormous but the work in front of him was small enough to act on.
There is good evidence behind that approach. A long-running body of goal-setting research has found that specific, challenging goals tend to outperform vague “do your best” instructions, particularly when feedback is part of the process. Research on deliberate practice makes a similar distinction: improvement comes from focused work on a particular task, repeated performance, feedback and evaluation, not simply doing the same job for longer. A PubMed overview of deliberate practice summarises those elements clearly.
In marketing, “improve the content”, “get better at AI” or “increase engagement” are not practice goals. They are aspirations.
A useful improvement goal is narrow enough that the team can tell what it is changing. For example, you might spend the next four sales emails improving the first two lines so the customer problem is clearer, or use the next five campaign reviews to get better at separating a weak message from a weak channel. If AI is the skill, you might build one repeatable brief-to-first-draft workflow and measure where human editing is still doing most of the work.
Failure is useful only when it changes the next attempt
Brady told a simple story about good decisions coming from experience, and experience often coming from bad decisions. It is a familiar idea, but there is an important bit businesses sometimes skip: a mistake does not automatically turn into experience.
A campaign that underperforms does not magically make the team wiser. A poor sales call does not improve the next one by itself. A piece of AI-generated copy that sounds like it was assembled by a committee is not a learning experience unless someone works out why it happened and changes the process.
After a meaningful piece of work, ask four things:
- What did we expect to happen?
- What actually happened?
- What is the most likely reason for the gap?
- What will we do differently on the next rep?
Keep that review small enough to happen. Not every email needs a two-hour post-mortem and twelve people in a meeting room. The aim is to create a habit of learning, not an administrative punishment for trying something.
The AI learning curve is exactly the point
Towards the end of the UNBOUND session, Halligan put the question directly into the context of AI. Everyone was leaving the conference with a list of things they wanted to build, but many would get back to work, try something unfamiliar, “skin their knees” and stop.
Brady’s answer was refreshingly uncomplicated: start. Expect the early attempts to be rough. Improve something. Then do it again.
That may have been some of the most useful AI advice we heard all week.
At STORY22, that matches our own experience with AI. We have spent years testing AI in real marketing workflows, training teams and working out where it saves time and where human judgement still needs to stay firmly in the loop. The biggest improvement has come from better briefs, better context, clearer standards, repeated use and a much more disciplined review process.
AI can shorten the distance between one attempt and the next but it does not remove the learning curve. In some ways it makes the need for judgement more obvious, because you can now create bad work at extraordinary speed.
A simple marketing practice loop
If you want the Brady lesson without adopting the lifestyle of an elite professional athlete, use this five-step loop on one marketing skill at a time:
This is where marketing performance starts to compound. Because every meaningful attempt has a job beyond its immediate result: it should make the next one better.
You do not need Tom Brady’s obsession. You do need a practice system.
Most marketing teams do not need to work harder. They need a better way to learn from the work they are already doing.
The lesson we brought back from Brady’s session was not “want it more”. It was much more practical than that. Give people reps before the pressure is high. Make improvement specific. Get feedback quickly. Let mistakes teach you something. Keep going long enough for the small gains to show up in the work.
That is especially important now. AI has made it possible to try more things, more quickly, but more activity is not the same as more capability. The teams that get genuinely better will be the ones that use the extra speed to create a tighter learning loop rather than simply producing more stuff.
When the big moment arrives, whether that is a pitch, a product launch, a major campaign or a client presentation, the aim is not to suddenly become brilliant under pressure. It is to recognise the situation because you have already done the work.
If your marketing is busy but the team is not getting better
STORY22 helps businesses build the strategy, messaging, processes and review rhythms that turn marketing activity into something the team can learn from and improve. See what a strong marketing strategy needs in 2026, or book a free 30-minute strategy call if you want us to help identify where the real performance gap sits.
FAQs
How can a marketing team practise without wasting time?
Choose one real skill and create a lower-risk version of the work. Rehearse a pitch internally, test messaging with a small audience, review a draft AI workflow before it touches customer-facing work, or role-play common sales objections. Practice should be close enough to the real task to teach you something, but cheap enough that mistakes are useful rather than expensive.
How often should marketing performance be reviewed?
Review fast-moving activity often enough to influence the next attempt. A campaign team might do a short weekly learning review, while the wider marketing strategy may need a more formal monthly or quarterly check. The useful rhythm is the one that lets feedback change the work before the same mistake is repeated.
What is the difference between marketing experimentation and random testing?
An experiment starts with a clear question and a reason for changing one thing. Random testing changes several variables, waits for a result and then tries to reverse-engineer a lesson. The aim is not to test constantly. It is to reduce uncertainty about a decision that matters.
Does this approach apply to AI marketing?
Yes, particularly because AI makes it easy to produce a lot of output before a team has developed the judgement to assess it. Start with one workflow, define the quality standard, review the output, improve the briefing and repeat. STORY22’s practical AI marketing training is built around using AI as a repeatable working skill, not a one-off demo.
What should a marketing improvement plan include?
Keep it small. Name the skill you want to improve, define what better looks like, decide how the team will practise it, identify the feedback you will use and agree when you will review the next attempt. If the plan cannot influence what happens next week, it is probably too abstract.
About STORY22
STORY22 is a full-service international marketing agency based in the UK. We help organisations make clearer marketing decisions by connecting strategy, customer-centred messaging, digital marketing, AI, content and measurement. Our work is designed to give teams a practical system they can use, review and improve, rather than another plan that looks good in a meeting and disappears into a folder.