Customer Playbook
Customer archetypes and their operational expectations.
Browse 56 customer profiles and learn how their requirements affect capacity, reliability, and reputation.
-
Ada's platform bids on ad impressions in real time. The auction closes in 50 milliseconds whether her servers answered or not.
-
Ade runs a digital agency hosting a dozen client sites. Quick to point the finger at the host when a client complains, and always shopping for a cheaper deal.
-
Trains models. Needs GPUs. There are never enough GPUs.
-
Quiet most of the time, brutal during ETL windows. If you've over-packed his rack with other tenants, he'll notice.
-
Twenty percent of your revenue, in one customer. Pays cheap because he's huge. Will personally fly to your DC to give you stick about uptime.
-
Avi's whole product is an API. Latency is the value proposition, so every extra millisecond is a developer quietly switching to a competitor.
-
Sends huge backup blobs nightly. Reads them only when something terrible has happened to the customer. Mission durability — losing data here ends careers.
-
Dana's whole livelihood runs on these VMs. She panics about uptime and asks architecture questions as support tickets — treat her well and she's your best teacher.
-
Bailey's team spins up a CI cluster, runs it hot for an hour, and tears it down. Pays by the second, comes and goes, never sentimental about a provider.
-
Brandt runs the bank's core ledger on the mainframe and nothing else. Five-nines or bust, PCI, and a year to sign. The most valuable — and least forgiving — tenant you'll ever…
-
Chad pays well and on time, right up until the day he doesn't. His workloads attract DDoS attention you'll have to scrub, and he'll churn the instant something's slow.
-
Dani's whole product depends on her database staying fast. She'll pay premium for managed Postgres with backups, but expects no surprises.
-
Dan ships microservices and wants Kubernetes as a service. He pays a flat fee for the control plane and an hourly rate per worker. Take down a worker and his cluster degrades;…
-
Bex's bot lives in a bunch of servers. Barely any load, talks in lowercase and emoji, and is delighted when things just work.
-
Eleanor runs an online shop. She's chill 11 months a year. December she becomes the most stressed person on Earth, and she expects you to be ready.
-
Growing fast, demanding fast. If you can keep up he'll grow into a Gold/Platinum customer. If you can't, he'll bring his runway elsewhere.
-
Ed's platform runs lessons for schools. Dead on weekends, slammed during class hours, and an outage mid-lesson is a hundred angry teachers.
-
Enid is calm money. She doesn't ask for much, just steady performance and a clean SOC2 report. Lose either and she's gone — and her replacement won't sign with you for years.
-
Priya runs a careful enterprise pilot. Demands SOC2 and audit logs, takes weeks to sign, then sticks for years — unless you fail an audit, which ends it on the spot.
-
Like Enid but on dedicated VMs and stricter regulation. Needs HIPAA. The kind of customer that takes a year to sign and a tweet to lose.
-
Etta sells handmade goods from a tiny shop. Calm most days, but a promo brings a rush — and the checkout cannot be the thing that fails.
-
Fred handles payments and balances. SOC2 is the price of entry, uptime is existential, and a single compliance failure ends the relationship.
-
Faye's app tracks workouts. Two rushes a day — before and after work — plus a January resolution wave that dwarfs everything.
-
Fiona juggles a handful of client sites. Semi-technical and stretched thin — when a client complains, the host hears about it first.
-
Gabe runs a game server for his friends. He needs RAM and low latency; the moment the server stutters, his whole crew is in his DMs.
-
Saturday evenings his game is full of players. Tuesday afternoons it's empty. Latency-sensitive on weekends, doesn't care on weekdays.
-
Gita's lab processes genome datasets on grant money. Enormous storage and GPU batch jobs — and a refreshing tolerance for queue time.
-
The prestige tier. Six months to onboard, air-gapped, audited every year — but once you're in, they never leave. Lose them only by failing compliance.
-
Hana's platform holds patient data. HIPAA is mandatory, audit logs are always on, and a compliance slip is a one-strike departure.
-
Harold runs his blog out of curiosity. He pays a few dollars a month and will leave the moment his bill jumps.
-
Hank runs Nextcloud and Home Assistant around the clock. Deeply technical, tinkers constantly, and resents paying for anything he could self-host.
-
Isabella's running her one-person SaaS and watching every metric. One bad night and she's tweeting about you.
-
Aoki runs decades of policy math in overnight batches on the mainframe. Quiet, audited annually, and gone if a batch window slips.
-
Ivan's product is a fleet of devices that constantly report in. Tiny payloads, huge connection count, and it grows with every unit sold.
-
Carl was told he 'needs a website'. He doesn't know what a server is, panics at jargon, and one baffling bill will send him to a website builder.
-
Distributes media files. Wants high cache hit rates because misses are expensive for both of you.
-
Devi's shop knows exactly what it wants and pays for stability. The prestige tier of the VM market — keep them happy and they vouch for you.
-
Modern SaaS company that needs SOC2, a hot Postgres replica, and Mission-durability backups. The bundle is the product.
-
Mike runs the backend for a consumer app. Traffic follows the clock, then a push notification sends everyone back in at once.
-
Nadia writes a newsletter to a growing list. Most days are calm; send day is all that matters, and it has to go out clean.
-
Mo maintains a popular open-source project and hosts its docs. No budget, but a megaphone in dev circles — keep Mo happy and word travels.
-
Stores customer photos. Cares deeply about durability — losing one wedding album is a viral horror story.
-
Pete's product is a database, period. He needs a hot replica, Mission-tier backups, and zero excuses.
-
Rita runs a small online store with real revenue. Calm most days; a flash sale makes the checkout path the entire business for an hour.
-
Fen builds firmware for RISC-V devices, and the build farm has to match the target chip. Small but sticky — as long as you have the right hosts.
-
Rao's lab does open-hardware research and only runs on RISC-V. Patient and grant-funded, but they need matching silicon — no RISC-V hosts, no deal.
-
Sasha's company is growing fast and demanding faster. Keep up and she compounds into your biggest tenant; stumble once and she's gone to a hyperscaler.
-
Sid builds ambitious things on weekends and blames the cloud when they break at 2am. Forgiving on uptime, fickle on attention. Might go viral once.
-
Most days he's normal. Some days he's viral. When that happens, you're handling 3000 req/s for 10 minutes and your fabric saturates.
-
Olly runs his own stream overlay and chat tools. Hype on during the evening broadcast; the overlay has to keep up live or chat notices.
-
Steph is hosting a class project to get a grade. She's up at 2am, asks great beginner questions, and will forgive almost anything — it isn't production.
-
Tess's firm trades on speed. Latency is the entire product — she pays whatever it costs, tolerates nothing, and measures you in microseconds.
-
Vic's studio renders effects shots. Quiet between projects, then a deadline lands and they want every GPU you have, immediately.
-
Like Sven but lives on VMs instead of functions. Wants 2-6 boxes that grow over time.
-
Runs a small hosting biz behind your load balancer. Needs a public IP that stays stable.
-
Runs a real product on real hardware. Needs four services to work together; if any of them fails, her users see a broken site.