The Serve-First Client Interview
Before you start the job, ask twenty customers what they actually want from you.
- Difficulty
- Starter
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 90%
When Gurley arrived on Wall Street as a new sell-side analyst, he noticed the firm's salespeople owned the relationships with the buy-side accounts — Fidelity, Wellington, Texas Teachers. Rather than start publishing and hope, he asked each salesperson for one client who would spend 30 to 45 minutes with him, and he did roughly twenty of those interviews before he had even started the job. The framing was deliberately extractive of nothing: he had nothing for them and just wanted to know how he could serve them best. What came back contradicted the public image of the job. Clients told him they did not really need buy or sell recommendations; what they wanted was a point of view or a piece of analysis that caused them to think differently about a company or industry — work nobody else had done, that made them want to call him. He also asked each of them who did this well, then tried to befriend that person, which is how he ended up copying David Korus's weekly fax. Because Institutional Investor's All-America ranking is literally a poll of the buy side, knowing what they were looking for before he started was the whole game. He landed on the list.
Origin
Gurley developed this on arrival at his first Wall Street research job in the mid-1990s, after the firm introduced the analysts to the sales force and he realised the salespeople could broker direct access to the very buy-side clients who would later vote in the Institutional Investor rankings.
Core principles
- 01The people who rank you can tell you exactly what they want, and almost nobody asks them.
- 02Ask before you have anything to sell, so the conversation is not read as a pitch.
- 03The stated deliverable of a job is often not the valued deliverable — buy-side clients did not want buy/sell calls, they wanted analysis that made them think differently.
- 04Ask each interviewee who does this well, then go befriend that person; the referrals compound the research.
- 05Doing twenty of these before day one means you start with a map that peers acquire over years, if ever.
- 06The output is not information alone, it is a service posture that people remember and reward in polls.
How to run it
- 1
Map who owns the customer relationships
Identify the internal people who hold the relationship with each major account. In Gurley's case this was the sales force, one person per institution.
Pro tip These people are usually flattered to be asked and cost you nothing politically.
- 2
Request one interview per relationship owner
Ask each for a single client who will give you 30 to 45 minutes. One ask per person keeps the favour small and the yes rate high, and twenty relationship owners yields twenty interviews.
Pro tip Specify the length up front — a bounded ask converts far better than an open one.
Watch out Asking one person for twenty introductions burns the relationship; spread the ask.
- 3
Run the calls before you have anything to sell
Do the interviews before you start producing. State plainly that you are not there to have anything for them, you just want to know how to serve them best.
Pro tip Having nothing to sell is the credibility unlock, not a limitation.
Watch out Once you have a product, the same conversation becomes a sales call and the answers get polite.
- 4
Ask who already does it well and get introduced
Close every interview by asking which practitioner they rate highly. Then try to befriend that person and study their format directly.
Pro tip Copy the format, not the content — Gurley copied David Korus's weekly fax cadence.
- 5
Build to the stated need, not the assumed one
Synthesise the pattern across interviews and make that the core of your output. For Gurley the pattern was original analysis that changed how clients thought, not ratings changes.
Pro tip Write the pattern as a one-sentence mandate and test every piece of output against it.
Watch out If your synthesis matches what everyone already produces, you have interviewed too few people or asked too generically.
In the wild
Gurley asked each salesperson at his firm for one buy-side client who would talk to a new analyst for 30 to 45 minutes. He completed roughly twenty of those conversations before he had started the job, purely to learn what the clients wanted.
→ He learned the buy side wanted differentiated analysis rather than buy/sell calls, built to that, and made the Institutional Investor All-America research team — a ranking decided by a poll of those same clients.
As part of the interviews Gurley asked who did the job well. That surfaced David Korus, who published a weekly fax to clients. Gurley studied and then adopted the same weekly cadence after Korus left the field.
→ The weekly piece became the vehicle for his differentiated analysis and, later, for the distribution push that put him on the map.
Common mistakes
Waiting until you have something to show
The interview only works while you have nothing to sell. Delay it until you have a product and the conversation becomes a pitch, the answers become polite, and you learn nothing.
Assuming the job description is the value
Gurley's clients explicitly said the headline deliverable — the buy or sell recommendation — was not what they valued. Building to the official output is how you end up indistinguishable from everyone else in the seat.
Running two or three interviews and calling it research
The value came from volume — around twenty conversations — which is what let him see a pattern rather than one person's idiosyncratic preference.
Is it for you?
Best for
Anyone entering a role where an audience rates, ranks or renews them and where the real value driver is ambiguous.
Not ideal for
Roles with a single, unambiguous, mechanically measured output where preference discovery adds nothing.
From the transcript
“I decided to ask each salesperson is there one client that will spend 30 to 45 minutes with me as a new analyst and just…”
“20 of those interviews before I had started the job”
“I don't really need you to make this buy or sell recommendation like that's not what I what I would really get a huge benefit…”
From the episode
#651: Legendary Investor Bill Gurley on Investing Rules, Finding Outliers, Insights from Jeff Bezos and Howard Marks, Must-Read Books, Creating True Competitive Advantages, Open-Source Strategies, Adapting Mental Models to New Realities, and More