Where we start
The Partnership track is about the working relationship, not the numbers. Four sessions: tonight, the other side's world; then who owns which ask; then the front door, an intake template and a service menu; then how to say no well, and when to say yes. The running company is Kettle & Co, a constructed direct-to-consumer tea brand with a marketing lead, a brand manager, a BD lead and a data team of three. Every Kettle & Co person, date and number is invented, and the pages say so. The roles are real: almost every company that sells to consumers has these three, and almost every analyst serving them has met each of the complaints below.
★ What you walk out with tonight. A one-line answer to "what is this person judged on, and by when" for each of your stakeholders, the habit of reading an ask against their calendar rather than yours, a translation table for five words that cause most of the misunderstandings, and a stakeholder map you can put on the wall.
Three roles, three scoreboards
6 min liveMarketing, brand and BD are often treated as one customer of the data team. They are three, and they are judged on different things over different horizons. The marketing lead is judged this quarter, on what campaigns brought in. The brand manager is judged over years, on what people think of the brand, read in survey waves. The BD lead is judged on deals, over a cycle of months, in rooms the data team never enters. The same analysis lands differently on each of them.
LiveRead the scoreboard before you read the ask4 min▶
The marketing lead owns a budget and a calendar, and is judged on what the budget brought in, usually as revenue, new customers, cost per acquisition (CAC) and return on ad spend (ROAS). Their asks are short-fused because their dates are fixed. When they ask "is the campaign working?", the question underneath is "do I keep spending on Monday?".
The brand manager owns how the brand is seen, and is judged on awareness, consideration and consistency, read in survey waves months apart. Their asks are slower and more defensive: "show me this promotion is not hurting us" is a common one. A dashboard of weekly clicks rarely helps them, and a careful read of a tracker wave with its uncertainty often does.
The BD lead owns partnerships and channels: retailers, B2B accounts, marketplaces, distributors. They are judged on deals, in meetings the data team never attends. Their asks arrive as "I need numbers for Thursday", and what they need is one page they can defend across a table from a buyer.
Self-studyWhat a data team is usually measured on, and why that matters here3 min▶
Data teams are often measured on throughput: tickets closed, dashboards delivered, uptime. None of those appear on the three scoreboards above. A data team that closes forty tickets in a month in which the marketing lead missed a launch has not helped, by the only measure the marketing lead has. The fix is not to adopt their scoreboard as yours, which is how analysts end up owning strategy (Partnership session 2). It is to know their scoreboard well enough that every deliverable says which line of it the work serves.
The dashboard nobody opened. A data team ships a beautiful weekly brand dashboard: impressions, followers, engagement rate, all trending up. The brand manager never opens it after week two. When asked, she says the numbers do not tell her whether people think better of the brand, which is what she is judged on, and that she gets that once a quarter from the tracker anyway. Constructed example; the mismatch is common.
Their calendar was fixed months ago
5 min liveMost marketing dates are set long before the data team hears about them: the autumn launch, the budget round, Black Friday, the Christmas send. An ask that lands "too late" usually landed exactly when the requester's own plan reached the point of needing evidence. A data team that plans its capacity around its own sprint and ignores the marketing calendar will be surprised by the same crunch four times a year.
LiveThree habits that make the calendar work for you3 min▶
Get the calendar. Ask the marketing lead for the campaign calendar and the BD lead for the deal pipeline dates, once a quarter. Put both into the data team's planning, so the weeks before each fixed date carry less internal work.
Pre-agree the readout. For every campaign on the calendar, agree the success metric and the readout date before launch. A readout negotiated after the campaign ends turns into a search for a metric that looks good.
Name the crunch. When two fixed dates collide, as Black Friday and the Christmas send do, say so in week one, with the capacity you have. That conversation is Partnership session 4.
Self-studyWhy "plan ahead" is not the answer you can give them2 min▶
Telling a marketing lead to send asks earlier rarely works, because their own inputs arrive late too: a retailer moves a slot, a creator drops out, the budget round cuts a channel. The data team is in a better position to absorb timing than to change it. What it can change is how predictable the crunch is, and how small each ask is when it lands, which is what an intake template (Partnership session 3) does.
Five words that mean two things
4 min liveMost "the numbers don't match" arguments between marketing and data are not arithmetic. They are one word with two definitions. The platform's conversion is any goal event the platform counts; the data team's conversion is usually an order in the company's own database. Both are right. Neither is the other.
LiveHow to correct a definition without correcting a person3 min▶
"That's not a real conversion" is accurate and lands as an insult. "Meta counts 412 conversions; our order table has 268 orders from the same campaign; here is what each one counts" is accurate and lands as help. The habit is to show both numbers side by side with their definitions, and let the marketer choose which one their decision needs. Usually they choose yours, and they remember that you did not make them feel foolish. (Both figures in this paragraph are invented for the illustration.)
Angles sessions 2 and 3 take each channel's own metrics in turn: what the platform reports, which number looks good and decides nothing, and where the data team adds value.
Self-studyBuild a glossary they own2 min▶
A glossary written by data and imposed on marketing is ignored. A glossary written with the marketing lead, in their words, with the data definition next to each, gets used, because it is theirs. Keep it to the twenty terms that actually appear on their slides. Every term carries an owner, which for platform metrics is the platform, and a note on how it differs from the company's own count.
Two numbers in one meeting. A marketing lead presents a campaign with "1,200 leads" from the platform; the finance partner has "300 new customers" from the order system. Twenty minutes go on whose number is wrong. Neither is: one counts form fills, the other counts first orders, and about a quarter of the leads bought. Put both on one slide with their definitions and the meeting is about the conversion rate, which is the question that mattered. Constructed example; the numbers are invented.
The Kettle & Co stakeholder map
★ 22 min · everyone works it throughNo code. Six steps that turn "marketing" into five named people with five scoreboards, placed on one map, with one thing you can bring each of them next month. Work it for Kettle & Co first, then for your own company. Every placement on the Kettle & Co map is a constructed judgement, and yours will differ.
Step 1 · Name the people, not the department. "Marketing" is not a stakeholder. Write each person who sends the data team asks, with their role. Kettle & Co has five: the marketing lead, the brand manager, the BD lead, the finance partner and the ops lead.
Step 2 · Write each scoreboard in one line. What are they judged on, and by when? If you do not know, that is the first question for your next one-to-one with them.
Step 3 · Write the fear behind their urgent asks. The saffron band from Part 1. It is usually one sentence, and it predicts the asks better than their job title does.
Step 4 · Place them on the map. Two questions per person: how much of the outcome do they carry, and how much do their decisions depend on data? Rough is fine. Kettle & Co's marketing lead goes top right; its ops lead bottom left.
Step 5 · Pull their calendar onto yours. For the next quarter, list every fixed date each person has, and mark the two weeks before each as a crunch window in your own plan.
Step 6 · One thing to bring each of them next month. Not a dashboard. One deliverable that serves a line on their scoreboard, with a date. This line becomes the start of your service menu in Partnership session 3.
Keep the map on the wall. Redraw it every quarter. People move, and the top-left corner, the stakeholder who carries the outcome and leans on data less, is the one who most often ends up deciding on anecdote because nobody brought them anything.
Before Partnership session 2
30 min- Finish your own stakeholder map with real names, a one-line scoreboard and a fear for each.
- Get one real calendar. Ask your marketing lead or BD lead for next quarter's fixed dates, and mark the crunch windows in your team's plan.
- Start the translation table with the five words from Part 3 and two of your own, each with the platform's definition and your team's.
- List the last ten asks you received, with who sent them. Partnership session 2 tags each one by the knowledge it needs.
Three questions
Clear all three and Partnership session 1 gets stamped in your passport.
1 · The brand manager ignores a weekly dashboard of impressions and followers. What is the most likely reason?
Read the scoreboard before the ask. A deliverable that serves no line on her scoreboard will not be used, however accurate it is.
2 · Asks from marketing arrive in bursts that the data team calls "last minute". What does Part 2 suggest?
The dates were fixed months earlier and the asks arrive when the plan needs evidence. The data team can rarely change the timing; it can make the crunch predictable and the asks smaller.
3 · Meta reports 412 conversions for a campaign and the order table shows 268 orders. What is the useful first move?
One word, two definitions, both correct for what they count. Correct the definition, not the person. (The two figures are invented for the question.)
What this session teaches, and where it came from
Every fact on this page, and its verification tier, is recorded in the course's source map.
Partnership session 1 cheat sheet · pin this
You know whose shoes you are standing in. Partnership session 2 asks whose job each ask is: every ask tagged by the knowledge it needs, a RACI for 24 of them, and a triage desk where saying yes to everything and saying no to everything fail in different columns.