Andrey Koh
Andrew Koch
B2B SaaS Product Management, Product Strategy & Analytics
Name, title and affiliations are given by the model’s author; avva has not verified them. Report this model
Before your AI advises you on B2B SaaS product Management, product strategy & analytics, it reads the decisions Andrey actually made, the reasons where they recorded one, and where Andrey would push back.
Use this model for
- KPI and metric definition
- Experiment and rollout readiness
- Activation and retention measurement
- Pricing and packaging
- Scope and enterprise customization
- Checkout and fulfilment rules
Built from 30 approved decisions across documents. Updated November 4, 2026.
Add this link to Claude, ChatGPT, or Cursor. Your AI then checks Andrey’s decisions before it answers you. The answer is still your AI’s.
https://On a phone? Email the link to yourself and add it on a computer.
What your AI sends is used for the answer and not stored. Andrey is not told.
- Try it free
- No account, up to 40 requests a day, per model. It returns a smaller part of the model: every stop rule and question, their approach and defaults, the criteria they put first, and five of their decisions.
- Full access
- Also free: sign in once, then up to 1,000 requests a day, per model. It returns the complete published model, on a connection you can revoke from any device.
An example AI answer
An example approved by Andrey, with the published decision behind it. Your AI writes the answer to your own question.
Example inside your AI app: a question Andrey chose, and the answer Andrey approved
A large prospect wants a calculator that is not on our roadmap. Sales wants to promise it in the core product this week to close the deal. Should we?
Checked Andrey’s Decision Model
Andrey Koh would refuse to put that calculator into the core on the strength of one deal. He would first describe the job, check whether other clients need the same thing, then choose a configurable module, paid custom work, or an explicit decline — so one contract does not quietly rewrite platform strategy.
Why, and what would change it
Why. A single deal does not prove mass need; a unique branch stays for a long time.
What would change this. Repeat demand from other clients after the job is described — then a configurable module or paid custom can proceed.
Outside this model. Exact pricing of paid custom, legal terms of the deal, or whether this prospect's calculator maps to an existing module.
Copy an example prompt4 patterns with a gap to fill, once you are connected
Before you commit
A full pass over a plan or a draft, with the changes it needs.
Use Andrey Koh's Decision Model on avva before we lock this in. What we're deciding: <paste your plan, doc, or diff here> Run their review, follow the steps it returns, and give me the verdict and any required changes before you continue.Choosing between options
A choice between two options, their stop rules checked first.
Use Andrey Koh's Decision Model on avva to compare these. Option A: <describe it> Option B: <describe it> Context: <what constrains the choice — experiment and rollout readiness> Check their stop rules first, then compare the options under their judgment, and tell me which one their model favors and why.A quick gate
A quick look for anything they would refuse outright.
Run this past the stop rules in Andrey Koh's Decision Model on avva before I send it. <paste the draft, plan, or change> Tell me only what they would refuse or flag, and why. If nothing trips, say so.Before you have an approach
The questions they ask before an approach gets locked in.
I'm starting on KPI and metric definition. Ask Andrey Koh's Decision Model on avva what they would want settled first. Situation: <two or three lines> Give me their questions, and flag the ones my situation does not answer yet.
Patterns with a gap to fill. None is an answer Andrey gave. The one worked answer on this page is the one they approved themselves; anything else here would have been written by us, under their name.
Connect Andrey’s Decision Model to your AI
Three steps. The first works now, with no key or account; the last makes your AI check Andrey’s model without being asked.
The link works in AI apps that can add a remote MCP (MCP is the open standard for plugging in outside tools). You set it up once.
A connector is an address your AI reads from: one line, nothing to install.
https://avva.chat/ mcp/ b2b- saas- product- management- product- strategy - Open Claude on the web or in the desktop app and go to Customize → Connectors.
- Press + next to Connectors and choose Add custom connector.
- Paste the address above and name it avva-b2b-saas-product-management-product-strategy.
- When Claude asks to allow avva, say yes.
Missing on a work account? Ask your admin to allow it. Claude Code and Cursor take a local config instead.
The link is added on the web or in the desktop app. Email this line to myself
The trial is enough to see how this expert reasons and whether their thinking fits your problem. Step 2 returns the rest.
Ask your first question. Send this to your AI and continue there.
Help me with a decision using this Decision Model on avva: "Andrey Koh" (model: "b2b-saas- product- management- product- strategy"). First ask what I need to decide, including my options and constraints. Then call this model's review tool and follow the returned procedure. If you cannot call it, tell me and stop. Base your recommendation on its published decisions and explain what supports it and what the model does not cover. Connection not working?
Check it worked. Send this in the chat once the link is added.
Ask Andrey Koh's Decision Model on avva what kinds of decision it covers, and summarize it for me in three lines.Look for a call to this model in your AI’s tool activity. If it is missing, check the setup steps above.
Not legal, medical, financial, or other professional advice. Terms of Use
About this model
Every rule below rests on decisions Andrey approved. Open any of these to check the grounding yourself.
How this model is groundedWhat it holds, and what Andrey weighs first
This model holds 30 decisions, 14 forces, 8 criteria, and 14 stop rules.
What Andrey weighs first
The expert set this order. 3 of 8 criteria are shown, the ones the model puts first. The other 5 reach your AI too, along with what each one looks for, the stop rules, and what changes Andrey’s answer.
Andrey’s usual answer changes under 6 named conditions. Those conditions and the 14 stop rules reach your AI with the rest.
See examples of approved decisions5 of 30, written from Andrey’s approved decisions
The first 5 in the record, in published order. They are not the ones nearest to your situation, and not a summary of the rest. Your AI does not copy a decision it reads here: it checks what decided that case against your own facts, and weighs it beside which pressures tended to settle Andrey’s cases. With a key or signed in, your AI receives all 30, ordered against your situation.
What decided itcontribution after variable costs
Official brand store that also sells via marketplaces. The team proposed site revenue and direct-channel share as the project goal.
Do not accept revenue or share alone as the goal. Measure channel contribution after variable costs and watch total brand economics in parallel. If a mechanic raises orders but cuts profit, it is not a success.
Why: Site sales growth can cannibalize marketplaces, raise logistics cost and reduce total brand profit.
What decided itshared core without fork sprawl · evidence before mechanism or claim
Corporate rollout talks for B2B SaaS. A large prospect asked for a calculator not on the roadmap; sales wanted to promise it in the core to close.
Do not put it in the core product immediately. First describe the job, check demand with other clients, then choose a configurable module, paid custom, or decline.
Why: A single deal does not prove mass need; a unique branch stays for a long time. One deal's revenue must not quietly rewrite platform strategy.
What decided ittrustworthy numbers before viz · buyer and buyer-side trust
Demo of an analytics cabinet to client leaders. New client, deal and revenue meant different things in CRM, accounting and the panel. Proposal: ship the report and clarify terms later.
Refuse. Build a metric dictionary with formulas, sources, refresh time, exclusions and an owner per definition. Agreed numbers first, visualization second.
Why: A pretty chart with ambiguous metrics destroys trust and only speeds fights about data.
What decided itoutcome metrics over activity
B2B SaaS for dealer networks after launching client cabinets. Team proposed admin MAU as the main proof of value.
Reject MAU as the target metric. Split usage vs outcome: client activation, time to first published storefront, share of companies regularly updating offers, retention of paying orgs. MAU stays diagnostic, not the goal.
Why: Staff can log in often because the process is painful, not because they get value.
What decided itability to fulfil the promise · buyer and buyer-side trust
Checkout for items not on own warehouse, fulfilment depending on supply. Simple path: take online payment for any catalog item immediately.
Do not take money until a manager confirms availability and fulfilability. For missing items: manager confirmation before payment; no partial shipment.
Why: Conversion to payment must not be bought with future cancellations and customer conflicts.
How Andrey takes up a decision5 steps in order, approved by Andrey
Step 1Name the metric unit and economic question first — contribution, churn type, activation definition — before accepting a chart, campaign win or KPI.
Backed by 5 decisions
Step 2Ask what evidence, segment cut or sample the claim rests on before approving personalization, experiment rollout or causal language.
Unless: first-party data and a measurable hypothesis already exist
Backed by 4 decisions
Step 3Freeze disputed fulfilment, document and regional rules with named owners before payment or release work proceeds.
When: order amount, documents or fulfilability are affected
Backed by 3 decisions
Step 4Unpack product meaning and repeat demand before estimating merge scope or putting unique work into the core.
When: scope, merge or enterprise custom is on the table
Unless: demand already repeats across clients
Backed by 3 decisions
Step 5Prefer a shared core and editor-clear offers over forks, warehouse-shaped catalogs and promo chrome.
Backed by 4 decisions
Where the order comes fromRead off the 30 decisions that say what won and what lost
- 30decisions Andrey approved, from documents
- 14pressures those decisions turned on, the ones Andrey decides under
- In groupswhich pressures tended to settle cases over which
Nobody ranks these pressures by hand. Where any order between them comes from is on For AI agents.
How Andrey decides, in summaryA summary, written from the decisions they approved
Consult this model when judging B2B SaaS and commerce analytics calls on KPIs, experiments, activation, pricing, scope, checkout promises and release readiness — especially when headline metrics, one-deal custom work or ship-now pressure conflict with contribution, defined units and fulfilability.
The paragraph above covers what the model holds. Andrey knows more than is in it.
What’s trustworthy hereWhat is checked, and what is only stated
- Identity: Name, title and affiliations are given by the model’s author; avva has not verified them. Judge by the record on this page and by what your AI receives.
- Scope: B2B SaaS Product Management, Product Strategy & Analytics (KPI and metric definition, Experiment and rollout readiness, Activation and retention measurement, Pricing and packaging, Scope and enterprise customization, Checkout and fulfilment rules, Catalog and offer clarity, Release and integration quality)
- Coverage: A solid set of decisions behind it that say why, with patterns that repeat across them.
- Privacy: Their cases and rules, written in English from decisions they approved. Their own AI takes out client and employer names, and they check every case; source quotes never publish.
- Whose answer it is: avva runs no AI of its own. Your AI reads this model and answers you, and it names where the model had nothing to say.
- Limits: Informational tooling for your AI, not professional advice. avva is provided as-is.
Know someone whose decisions you would want your AI to have? Send them to Studio. Publishing is free during beta.