Notes for agencies that punch above their weight
Field notes on agency economics, AI adoption and the craft of insight — written for independents, in-house teams and anyone tired of being out-gunned by the networks.
Lessons from the front line of going AI-native
Ten field notes built on one agency founder's 18 months of trying — and often failing — to make AI stick. Read them in order.
- AI adoption in agencies: document the real workflow before you automate
- AI-native agency: why speed isn't progress, and the 5 levels that are
- AI roles in agencies: stop making AI everyone's job
- Closed-loop AI for agencies: from eleven tools to one shared layer
- How to measure AI ROI in an agency: define success before you deploy
- AI metrics for agencies: why clients pay for results, not usage
- Agency AI strategy: don't build what Meta and Google will automate
- AI resistance in agencies is a standards problem, not a people problem
- AI readiness assessment for agencies: why your weakest dimension decides
- What to automate in an agency: three questions to answer first
What to automate in an agency: three questions to answer first
Three questions decide what an agency should automate: decision complexity, potential damage and measurable outcome. Start with the quick wins.
AI readiness assessment for agencies: why your weakest dimension decides
Your agency's AI readiness is set by its weakest dimension, not its strongest. A four-part assessment, and why another AI course rarely fixes the gap.
AI resistance in agencies is a standards problem, not a people problem
The people who push back on AI are often the ones with the highest standards. Make them approvers, not targets, and adoption stops feeding burnout.
Agency AI strategy: don't build what Meta and Google will automate
The platforms are automating what agencies once hired people for. Omnicom's AI platform shake-up shows the cost of building. Here's where to compete instead.
AI metrics for agencies: why clients pay for results, not usage
Clients don't care whether you used AI. They care whether the work is good. The AI metrics that matter as brands move away from paying for time.
How to measure AI ROI in an agency: define success before you deploy
AI that drafts a report in minutes can still cost you four extra hours of fixes. Here is how agencies define success before deploying AI, and count the full cost.
Closed-loop AI for agencies: from eleven tools to one shared layer
Fragmented tools leave agency AI guessing. One shared layer of memory, skills, workflow and approvals, fed by a closed loop, gives people and AI the full picture.
AI roles in agencies: stop making AI everyone's job
Giving every employee a code account and a build day leaves agencies with a mess. Three clear AI roles, builders, users and approvers, work better for people and for clients.
AI-native agency: why speed isn't progress, and the 5 levels that are
Racing to be AI-native without a finish line is the hare's strategy. The five levels of AI adoption show agencies where to climb, and where to stop.
AI adoption in agencies: document the real workflow before you automate
AI on a bad process just gives you a fast bad process. Before your agency automates anything, watch how the work really moves and find the one bottleneck worth fixing.
AI and consumer behaviour: why agencies need to rebuild their insight work
When people ask chatbots and algorithms curate their feeds, classic insight methods go stale. What strategists should do differently now – and what AI can't do for them.
The cold-start problem: how to personalise ads when you have no user data
New users, no history, few consents: how agencies solve the cold-start problem with personas built from real public conversations – instead of bought-in segments.
Christmas 2026: social listening as an early-warning system for agencies
Analyse it in January and Christmas is already over. How agencies use social listening to spot buying mood, price pressure and competitor promotions early.
AI tools for agencies: how MCP gets you out of the tool sprawl
Ten AI subscriptions, nothing joined up: why agencies with more tools often achieve less – and how the Model Context Protocol brings data to where your team already works.
Agency pricing models for AI: price it fairly, don't give away your margin
AI makes agency work faster, not free. Why the hourly rate is under pressure, what Mutabor's ‘M-Credit’ shows and how to price AI so that value, not time, is what counts.
Narrowing your agency's service offering: what to stop, and when
Agencies that narrow their offer grow faster. How to spot the service to stop – and decide on data rather than habit.
Measuring cross-platform engagement: time to retire the agency spreadsheet
Every platform counts engagement its own way. The method, metrics and tool criteria for reliable cross-platform measurement – no stitched-together spreadsheet required.
Imposter syndrome in agencies: rebuilding confidence in the AI era
Constant judgement, perfectionism, juniors wondering what they are for now that AI is here: how to swap bravado for evidence in agency life.
Research without respondents: are synthetic personas promise or mirage?
Research without human respondents is arriving in agencies. Useful for exploring, risky for concluding: how to use it without telling yourself stories.
Media agencies and AI: what the new era means for independents
AI is compressing execution and shifting media agencies' value towards advice. What that means for independents, and where to start.
Agency remuneration: how to prove your value with social data
The AACC wants to wean agencies off hourly billing. You still have to prove your value: here is how to turn social data into traceable proof.
Engagement metrics clients value (and the ones that are window dressing)
Likes don't pay the bills. How to connect engagement metrics to business results, with lessons from a real marketing mix modelling case.
Black Friday for agencies: automate reporting, win back time to optimise
Black Friday 2026 falls on 27 November. A six-week plan to get reporting off your team's desk and make faster calls when the peak hits.
B2B content marketing that doesn't send decision-makers to sleep
The Stagwell–ServiceNow deal shows B2B budgets are big. Why dull content costs you dearly, and how to create content buyers feel is worth their time.
AI and the end of the holding company model: lessons for independent agencies
WPP is set to cut another 1,000 roles and Gartner forecasts agency fees at 15% of marketing budgets by 2030. What the dismantling of the pyramid teaches independent agencies.
Fragmented media consumption: understanding your audience beyond dashboards
Every platform measures with its own ruler. Why adding up reports doesn't reveal your audience, and how to plan from personas built on real conversations.
Agency pitches in Brazil: the new rules reward those who bring evidence
Brazil's Abracom and ABA have updated their pitch recommendations. Transparent criteria favour agencies that can prove their case: here's how to build a category dossier that convinces.
Brand differentiation with AI: how not to sound like your competitors
AI can make every brand sound the same or help one stand out. The difference lies in what you use it for: writing the messages or mapping the category.
Agency burnout: fewer repetitive tasks, more time for creativity
Not all agency stress is equal. How to tell the pressure that gives the craft its meaning from the repetitive load that burns teams out, and which part you can automate.
Tariffs and marketing strategy: how to rebalance messages and budgets
Tariffs reshuffle entire categories. How to read their impact on consumers, rework messages market by market and move budget without sliding into paralysis.
Social listening for agencies: why the quarterly report is already too late
Consumer trends don’t wait for the quarterly report. How an agency sets up continuous social listening that turns real conversation into weekly decisions.
AI data analysis in marketing: conclusions your clients can verify
An AI conclusion nobody can verify is useless for decisions. Best practice for analysing data with AI in an agency, breaking down silos and answering “how do we know that?”.
AI for independent agencies: why they’re consolidating and what changes
Jungle combines four firms and SAMY buys in the US. AI is redefining what scale means for independent agencies, and merging is no longer the only way to get it.
Personalised advertising without losing trust: a guide for agencies
Personalisation no longer means stalking. How to tailor ads to niche communities, such as women’s football fans, from real understanding, and test what works before you spend.
AI media planning: from demographics to real consumer tensions
Demographics tell you who to buy, not why they’ll pay attention. How to plan media around your audience’s real tensions, with AI, without losing your judgement.
Media consumption data in Singapore: what to trust as networks shrink
Dentsu is cutting overseas entities and APAC is contracting. Here is how to find Singapore media consumption data you can trust, and how to read it without fooling yourself.
How AI hits each type of marketing agency: creative to influencer
AI does not reward scale, but it does not hit every agency type the same way either. A practical breakdown for creative, media, CRM, PR, influencer and social agencies.
AI content creation best practices: a quality framework for agencies
Cheap polish is everywhere. This framework puts human judgement at seven fixed gates, from brief to sign-off, so AI-assisted content keeps its voice and earns its media budget.
AI marketing case studies with real numbers, and what they leave out
Most AI marketing case studies report activity, not outcomes. Here are the ones with numbers you can check, what those numbers really mean, and a test for the rest.
AI-native marketing organisations: what clients still buy from agencies
Liquid Death's plan to run a sizeable media budget with two in-house people shows where client AI is heading. Here is what agencies stop selling, and what they sell instead.
Agency pitch strategy: win with a 48-hour category insight sprint
Lead with the client's category, not your credentials. A 48-hour sprint to find the angle no competitor has claimed, built from real ads and posts.
How to optimise marketing budget ROI when global campaigns underdeliver
Before cutting channels, check whether budget is backing creative that doesn't fit its market. A practical way to reallocate spend with evidence.
AI video production cost: why cheaper frames don't shrink the budget
AI makes every frame cheaper, so clients ask for more frames. Where the real savings sit in video production, and how agencies should price the work.
How to reach Gen Z consumers: plan from communities, not birth years
"Gen Z" is a birth-year bracket, not an audience. How to cluster real conversations into personas a creative team can actually brief from.
Retail media networks and agencies: who really owns the budget?
Retail media now spans retailer sites, social shops and creator content. Here is which agency type should own it, and why the category view decides.
Will AI replace marketing agencies? Why the maths rarely adds up
An AI subscription plus a hire looks cheaper than a retainer until you do the sums. Where AI genuinely replaces agency work, and where it frees people to do better work.
Agency consolidation or roster? What StarKist and ServiceNow reveal
StarKist chose one independent agency. ServiceNow chose a four-agency Stagwell team. Both deals reward the same thing, and it is not size.
Independent agency vs holding company: how independents compete on data
WPP is cutting 1,000 more roles and Omnicom has moved its AI platform team to a contractor. What the holdco data moat looks like now, and how independents compete.
Influencer marketing reporting: why it's the new agency battleground
Creator budgets are rising and creator agencies are consolidating. The agencies that win will be the ones whose influencer reports clients actually trust.
Audience research tools: what planners should automate first
WARC says fear of AI replacing strategists jumped 15 points this year. Here is what planners should automate first, and why freed time is where new thinking comes from.
Agentic AI advertising without the rogue fears: keep humans in the loop
AI agents are starting to plan and buy media. Here's how agencies keep humans in the loop, reassure nervous clients and stop brands from sounding like nobody.
Synthetic audiences are taking off. Here's how agencies keep them honest
Synthetic audiences are fast and persuasive, which is exactly why they need checking. A practical guide to using them without putting fiction in front of clients.
The agency sprint model: how to move fast without breaking strategy
Agencies are selling five-day sprints instead of retainers. Here's how to move that fast without the strategy falling apart, and which chores to hand to AI agents.
Creative pre-testing: how to end the creativity vs performance war
Creativity and performance only feel like enemies because the data arrives too late. Pre-test creative before spend and the evidence starts protecting bold ideas instead of killing them.
AI marketing tools for agencies: build a stack, not a shopping list
Most agencies don't have an AI stack. They have a pile of clever tools that don't talk to each other. Here's how to judge AI marketing tools on the thing that matters: integration.
Brand differentiation strategy: finding the angle your category ignores
Most categories look the same because brands copy each other. Use category intelligence as a map of the herd, not a recipe, and the unclaimed angle becomes visible.
Marketing dashboard overload: why agencies are swapping logins for questions
Only one in three marketers is happy with their martech. The fix isn't another dashboard. It's starting with the decision, asking a sharp question and insisting on evidence.
How to prove media ROI to clients: why MMM is back and what to show
Marketing mix modelling is back because clients want proof beyond last click. Here is the evidence stack agencies need to show media ROI, and the piece MMM cannot supply.
Outcome-based agency remuneration: the metrics beyond ROAS and CPA
Brands are abandoning hourly agency fees, but ROAS and CPA make poor foundations for outcome-based pay. Here are the metrics that hold up in a fee negotiation.
AI impact on agency jobs: how to rebuild the junior career ladder
Gartner and Digiday say AI is hollowing out entry-level agency work. Cutting the graduate pipeline is the wrong answer. Redesigning the junior role is the right one.