San Antonio 2022: themes of the day

On Monday 14th November 2022, the opening day of TMRE 2022, about 35 senior insight and analytics leaders, each invited individually, took a bus out of San Antonio to Knibbe Ranch, a heritage ranch in the Texas Hill Country about 28 miles north of the city. It rained all day. The day was built for conversation rather than presentation: introductions on the bus, short panels and talks each followed by a microphone passed around an open-air barn, a workshop, a fireside conversation and, later, roundtables and a campfire.

The day ran under the Chatham House rule: take the learning home, but not the name or the company attached to it. These notes cover the morning and the early afternoon, up to the start of the roundtables; the later sessions are not covered. They pool the day's thinking by theme, not by session, and carry no participant names, participant companies or sponsor names. Each view is framed as it was held in the room, by one participant or by several where several held it.

I. Honesty was the price of entry

The host opened with two rules. The first was the Chatham House rule, which one participant summed up as "what happens in Vegas stays in Vegas, sort of". The host tightened it: the story can leave the ranch, but not the name. Telling colleagues that an organisation faced a challenge and solved it in a particular way is fine. Saying who it was is not. Honesty, he said, would make or break the day.

The second rule was participation. At other events he runs, he said, people who only listen are not invited back. Here the sanction was softer and the principle the same: if half the room absorbs without giving, the discussion dies. The flip side of learning is sharing.

Asked what they wanted from the day, participants answered in three ways. Several wanted to learn how peers outside their own company and industry approach the same problems. Some ran small or junior teams with no one inside the business to sanity-check their thinking. And one wanted to raise the hard topics that are difficult to discuss internally: setting next year's budget with a recession looming, and new privacy rules that would shrink the data available.

A note of confidence ran through the round. One participant recalled years of the profession asking whether it would survive, and argued that the previous three years had shown how important it was. Several described the function as more central than it had ever been.

The host explained why he cared about the format. Years earlier, in another field, he had watched a small off-the-record group of five or ten people grow to dozens, sharing honestly about where they were falling behind on data. He believes that group changed how its whole field invested in insight.

So what for your team

• Share unresolved problems with a peer group as well as successes.
• Give as much as you take, and encourage everyone to contribute.

II. How a seat at the table is earned

One panellist described a team so central to the business that being in the room was assumed, not requested. The host asked who else was in that position, and how they got there.

The first answers pointed outwards. A new chief executive or chief strategy officer who believed in insight. A whole-company transformation that put the consumer at the centre. A pandemic that left leaders desperate for answers, so that one team was asked to answer "every single question" without the resources to do it.

The host pushed back. He suspected the room was being humble, crediting change at the top or chaos in the market for success that people had fought for. He asked for the battle scars instead.

The answers that followed were about hard, unglamorous work. One leader described acting as an internal consultant, helping teams with projects that were not theirs, and bringing stakeholders into research sessions so they heard customers first hand and became believers. Patience mattered. In one meeting a senior colleague admitted that, when the team first raised a finding eighteen months earlier, "we didn't believe you at all".

Another leader, building a research function from nothing, stressed showing value quickly while teaching colleagues how to use research, and prioritising ruthlessly so that the team is heard.

A third view was about restraint. Trust is built piece by piece, through a long run of small, well-judged contributions. That means humility about when to share what you know, and sometimes letting a leader believe the idea was theirs.

A fourth was about confidence. Researchers, one participant argued, walk into senior meetings thinking they do not belong, when nobody else in the room thinks that of themselves. The job is to grow the business, and selling research to a company that already pays for it is beside the point. Be humble, but be proud of what you do and stop second-guessing it.

Success brings its own problem. One leader who had moved a team "from order takers to strategic partners" found that everyone in the company then began calling their own data "insight". The new job was teaching the organisation the difference between information and an insight.

The room also disagreed with itself. At one point a participant challenged some of the combative language used during the day, such as talk of weaponising data. To them it felt male, and they doubted younger colleagues responded to it. They wanted the same strength in different words. The host thanked them and asked for more dissent through the rest of the day.

So what for your team

• When you describe how you won influence, name what your team did as well as what changed around it.
• Bring stakeholders into the research. People who hear customers themselves defend the findings.
• Expect to be early and disbelieved. Keep the evidence, and keep showing it.
• Once you are in demand, spend time teaching what an insight is, or everything will be called one.

III. Fewer pages, sharper questions

A repeated theme was the need to edit. Senior leaders want the few slides that make a decision easier, not pages of context, though the team must still know the evidence behind them, because the detailed question will come.

Others described the same discipline in different forms. One leader described insisting on a single page (the context, the highlights, the decision to be made and the next steps), and said that senior stakeholders want data access for their teams but, for themselves, want to be told the answer. Editing was also named as the core skill to grow in young analysts: in a world rich in data and starved of insight, the job is to pull out the few points that drive a decision.

Editing starts before the data exists. One panellist's team will not start a project until the requester can state the question, a hypothesis and the action they expect to take on the answer. Without that, they said, teams end up producing insights for questions nobody asked. Another recalled a mentor's advice that most of the effort in any project belongs to defining the problem and bringing the organisation along. The same logic applies to surveys. Overlong questionnaires produce bored respondents and weak data, and participants suggested removing any question that would not inform an action.

Stakeholders rarely make this easy. One leader said they are lucky to get more than a sentence or two from a requester. Their answer is a one-page commissioning note, and a habit: read what you send as the recipient would, and ask what you would do with it. Sitting in the business's strategy meetings helps too. Teams that hear the problems early are rarely surprised by the request.

Data democratisation drew a sceptical response. In theory everyone should have access. In practice, several said, it raises the value of context, and too often produces dashboards nobody uses. One leader called it, in some ways, a failed experiment, because the questions come back to the insight team anyway. The fixes offered were practical: organise data around what each kind of user actually needs, tell the same data differently to different executives, and send a short weekly note with the one thing people need to know.

So what for your team

• Make every request state its question, its hypothesis and the action it will drive. Do not start without them.
• Write the one page first, then decide what evidence it needs.
• Teach editing as a craft, in questionnaires as well as in reports.
• Before widening access to data, decide who will use it and what they need from it.

IV. Planning when the past stops predicting

Two panellists described a business climate in which history had stopped being a guide. One panellist noted that each of the previous six quarters had been different: supply chain problems in one, weather in another, inflation in a third. Year-on-year comparisons, even two- and three-year stacks, no longer explained much, so business partners were knocking on the insight team's door precisely because the usual data was not helping. The job, that panellist said, was to be ready to fill the gaps and to see around corners.

Methods had to move too. A tracker that would once have been kept consistent for years might now need changing week by week. Another panellist described the tension every researcher knows: protect the quality of the work and risk not being invited to the next conversation, or move at the speed of the decision and risk the quality. The answer offered was a mindset of agility without abandoning consistency.

On money, one leader put it simply: you need analytics in good times, and you really need analytics in bad times. That leader was still hiring, but argued for reviewing data spending against the value it demonstrably delivers. Big technology bets had become almost impossible to sell as a single grand promise of a box that would answer every question. What still sells is a phased plan with small, visible wins along the way.

One participant made the opposite point about predictability. Some events arrive on the same date every year, yet the requests for insight about them still come as a rush. Much of the pressure for speed is self-inflicted, and anticipating what partners will need is the cure.

So what for your team

• When the past stops predicting, say so plainly, and offer the business a way to read the present instead.
• Decide in advance where you will trade speed for quality, and where you will not.
• Sell data and technology investment in phases, with a visible win in each.
• Work out which requests are predictable, and answer them before they are asked.

V. Research and data science as one team

The relationship between research and analytics came up from both directions. One analytics leader argued that the two should be treated as one blended discipline: the job is to answer the business's questions with whatever tools fit. Another participant argued that bringing different disciplines together helps teams answer business questions faster, and that the friction between data science and research, which exists in many companies, can give way to partnership when each side sees the skills the other brings.

Hiring criteria, one participant argued, lag the work. Job adverts still screen for years of experience and coding languages, but senior marketers mostly want the business implications and where to look for growth, and that comes from connecting sources rather than from any one technique. So analytics teams need interpretation and communication skills alongside technical ones, including people who can make data legible.

So what for your team

• If research and analytics sit apart, put them on the same questions at the same time before you reorganise anything.
• Write job adverts for the work as it is now, which is more about connecting and explaining than about any one technique.
• Hire some people for their ability to make data legible, not only to model it.

VI. Hiring and growing people

Talent ran through the whole day, as it had through the questionnaires that participants filled in beforehand. Several leaders described hiring for potential over credentials. One assesses the ceiling: is this person smart, hungry and curious? Everything else can be taught. Another put it more bluntly: they can teach the maths, but not the ability to persuade a colleague to do something differently.

One leader supplemented interviews with a case study. Candidates, especially younger ones, receive a case study built on real data and present their findings. Some who interview brilliantly fail the case; some who interview poorly excel, and occasionally bring in evidence nobody asked for. Another agreed: young people are often polished interviewers, and the hard thing to find is someone who can build a persuasive story from the data.

On diversity, leaders described several practical levers. Set benchmarks for your team's recruitment against the rest of the organisation, and make progress a team goal. Recruit from internships and rotations, where you have seen how people work. Look beyond the usual pools and partner with colleges and community organisations. Once people are hired, help them find a community inside the company; one participant described an employee network that takes ownership of its members' pain points and carries them to leadership. Another made a point of meeting every new hire who shared their background, to share their own experience of the company.

A long discussion turned to introverts. One participant estimated that most of a typical insight and analytics team is introverted, and asked how to give those people a voice. One participant pushed back on the framing, saying many introverts communicate superbly, and the first speaker agreed that the issue is comfort rather than ability. Introversion is a spectrum, one participant argued, so start with self-awareness: where is someone comfortable, what is a stretch, what is the panic zone? Many people who dread a large room make the case brilliantly one to one. Build from there without asking them to change who they are.

The tactics offered were small and specific. A manager can open a meeting with a question that invites a junior colleague in, instead of leaving them to fight for airtime. On video calls, where an introvert is one box among many, use small breakout groups. Let junior team members set meeting agendas, to give them small moments of leading. Leaders who confess their own failures give their teams room to try. Remind people that no pitch happens only once: most ideas are sold several times before they land.

Two broader ideas also came up. One participant suggested looking for a unicorn team rather than a unicorn person: combine people who influence well with people who go deep technically, so that the team covers each person's weaknesses. Another argued that organisations need more than one route to the top, so that a specialist can progress without being forced to manage a team or stand on a stage.

So what for your team

• Consider testing candidates on a real case and a presentation; one leader found it more revealing than an interview.
• Hire for curiosity and the ability to persuade; teach the technique.
• Build each person's confidence where they are already strong, then widen it.
• Design the team, not the individual, to cover the skills you need, and create a senior path for specialists.

VII. Measuring what matters, and hearing who is missing

Several contributions questioned whether the numbers teams rely on measure what they think. One participant asked why brands built for the long term are judged on short-term returns. Another warned that what gets measured gets done, so choose the measure carefully.

One participant described posting a simple question to a neighbourhood social network and receiving replies only from the two extremes. That, they argued, is what social listening hears: loud minorities on each side, while a silent majority does not engage. Insight that leans on social data risks mistaking the extremes for the public. Another said their team's most credible moments came from triangulating what customers say with what they do. A third admitted to "a certain level of paranoia" before any big recommendation: is this built on solid numbers, or just on a report that happens to be finished? That participant said their team was finding it harder to reconcile syndicated and custom data.

Two participants raised the people that research misses. One works with a key group of users who are reluctant to appear on camera, so the team now meets them on the online platforms where they already gather, and explains plainly who the researchers are and what the research is for. The other raised a question that would grow much louder within a few years: the world's historical data carries its biases, so artificial intelligence built on top of it will perpetuate them, and in places make them worse. How, they asked, do you take the bias out of the data?

Empathy came up in a more literal form too. One leader argued that researchers should regularly use the products and services they study, as a customer would.

So what for your team

• Remember that social listening can overrepresent the most vocal people, and say so when you report it.
• Before a big recommendation, ask whether a second source agrees.
• Go to hard-to-reach groups where they already are, and explain why you want to hear from them.
• Use what you research. Regularly.

VIII. What the format taught

The day was designed so that nobody could simply consume it. Introductions began on the bus, with one question: if you were not doing this job, with unlimited resources and skills, what would you do? The host would be a mathematician. Architect and vet each came up more than once. After every panel and talk, the microphone went round the room.

The format had a cost, and participants named it. At lunch one told the host that every contribution opened a new idea and none was taken further, and suggested choosing one or two themes and going deep. The host agreed. Feedback after the event said the same: fewer topics, more time in small groups, and a chance to connect before the day. One post-event message judged the conversations between sessions more valuable than the stage talks.

The host's own impression was that the room started cagey and warmed up through the day, as people learned they were safe. His other reflection was simpler: even when a discussion reaches no solution, knowing that you are not alone with a problem is a help.

So what for your team

• Go deep on a few topics rather than skimming many. A full agenda is not the same as a valuable one.
• Start the conversation before the day, so that people arrive ready to be candid.

What happened next

Participants asked to stay in touch, and a contact list was circulated after everyone had the chance to opt out. The retreat ran again alongside the main conference in Colorado the following year, and later a virtual meet-up answered the request to connect between retreats.

Seen from today, one thing stands out. Artificial intelligence barely appeared in the questionnaires. In the sessions these notes cover it was mentioned in passing, as a skill on job adverts and a way of sorting large volumes of data, and only once as a subject in its own right: the warning about bias in historical data. ChatGPT was released to the public about a fortnight later.

If there is only one page

• Honesty is the price of entry. The story can travel; the name cannot.
• Influence is earned through visible, patient work, and by bringing stakeholders into the research. Then teach the business what an insight is.
• Edit ruthlessly: three slides, one page, fewer survey questions, and no project without a question, a hypothesis and an intended action.
• When history stops predicting, help the business read the present, and sell investment in small wins.
• Research and analytics work best as one team, and analytics needs broader backgrounds than its job adverts ask for.
• Hire for curiosity and persuasion, test with a real case, and build introverts' confidence where they are already strong.
• Be sceptical of the loudest data, and go looking for the people your research misses.