Four functions owned revenue. Which meant nobody did.
Keka, a $16.5M HR-tech company, was losing $2.24M a year to churn and discounting away 46% of list price, and nobody owned revenue end to end. As Head of Revenue Operations reporting to the CEO, I took ownership of the whole engine: the plan, the forecast, the deal desk, the retention motion, the data, and the stack. Here's what moved.
The problem
Revenue was split four ways. Sales owned bookings, marketing owned leads, customer success owned churn, finance owned the forecast. Every team was doing its job. No one owned the number end to end.
Which meant no one was accountable for what happened between the functions, and that is exactly where the money was going. Four numbers made the shape of the problem obvious.
Keka was losing $2.24 million a year to churn. Nobody owned the leak.
The leak nobody owned. $2.24M of revenue gone a year. Customer success found out a customer was leaving when the customer told them. No customer-360, no health signal, a 16% NPS response rate. Retention managed by anecdote.
A governance vacuum. Keka discounted away 46% of list price on average. Not on strategic deals, on average. A discount rate that high isn't a pricing decision; it's the absence of one.
A target, not a plan. No AOP, no forecasting cadence, no pipeline coverage model. A target is a number someone hopes for; a plan is the arithmetic that says how it happens. Three geos, two products, three segments. No arithmetic underneath any of it.
A stack nobody chose. $450,000 a year went into a GTM stack nobody had decided to buy. Tools accumulated the way they do when four functions each solve their own problem. The spend followed the fragmentation.
The diagnosis
I was hired to build a RevOps function. Keka didn't need a function. It needed someone to own the number.
A RevOps team that reports into sales optimises sales; one that reports into finance optimises reporting. Revenue leaked between the functions because every function was locally rational and nobody was globally accountable. The fix wasn't a team or a tool. It was a single owner, with the CEO's mandate, sitting above all four functions and answerable for the arc from lead to renewal.
The discount rate was the tell
A 46% average discount is a governance vacuum, not a pricing problem. Reps discount when nothing stops them and nobody measures it. You don't fix that with a pricing exercise. You fix it by putting a deal desk between the rep and the discount, and a number in front of the executive team every month.
You cannot fix retention you cannot see
With no customer-360 and a 16% NPS response rate, churn arrived as a surprise every single time. The instinct is to hire more customer success. The correct move is to build the signal first: a CS team without a health signal is just a team that apologises faster.
The stack was a symptom, not a cost problem
The $450,000 wasn't the disease. Four functions buying four sets of tools to solve four disconnected versions of the same problem. That was the disease. Rationalising spend without first unifying ownership saves money for one budget cycle, then rebuilds itself.
The CEO didn't need a report. He needed a counterpart
The most valuable thing I did was not a system. It was taking the revenue operating cadence off the CEO's desk (running the MBRs, QBRs and board reviews with the CFO and CRO) so revenue had a permanent, objective voice in the room that wasn't also carrying a quota.
What I built
Keka's first annual revenue plan, and the cadence to run it
There had never been one. I built the AOP across three geographies, two products and three segments, with capacity, pipeline and revenue planning underneath it, then the operating rhythm to run it: MBRs, QBRs and board revenue reviews, delivered on the CEO's behalf alongside the CFO and CRO.
A single view of the customer
A HubSpot Customer 360 with cohorted retention analysis: the first time churn and retention could be seen by segment, product and geography rather than inferred. NPS response rate went from 16% to 38%, turning a dead survey into an actual health signal.
A deal desk, and discount governance behind it
Discount approval moved from rep discretion to a governed workflow with tiered authority. Average discounting fell from 46% to 31%.
A pricing migration off grandfathered plans
Legacy customers sitting on old pricing were migrated onto current plans, generating $800,000 in expansion revenue from the existing base.
A GTM stack that somebody owned
The $450,000 stack was audited, consolidated and re-negotiated. Cost came down 40%, and what remained was integrated rather than adjacent.
New market and partner motions, built to be measured
A US market entry that reached $120,000 per quarter and grew 100% quarter over quarter; a partner programme where payout turnaround fell from 75 days to 40; dynamic lead routing; a GTM calendar; a multivariate compensation plan; and an outbound motion that lifted cold response from 5% to 8% at $0.08 per contact.
The outcome
Sixteen-month tenure. All metrics measured in-role.
He was the first person who owned the whole number instead of a slice of it, and told me things I didn't want to hear a quarter before I needed to. By the time he left, retention was something we could forecast.
The number we missed
Keka delivered 72% of its annual revenue target. I am not going to dress that up.
But the target had been set before the plan existed: the first annual revenue plan Keka had ever built, measured against a number chosen without one. What the plan and the forecasting cadence actually bought was not attainment. It was the ability to see the shortfall in month four instead of month eleven, and to make decisions about it while there was still most of a year left.
A forecast's job is not to make the number true. It's to tell you the truth about the number early enough to do something about it.
If your company is about to build its first real revenue plan, expect the same thing. The first plan does not fix the year. It tells you, honestly and early, what the year is going to be, and that is the more valuable of the two.
Why this matters for you
This is your problem if:
Revenue is owned by four functions, which means it is owned by nobody
Your average discount rate is a number no one tracks, and you suspect you would not like it
You find out a customer is churning when the cancellation email arrives
Your CEO asks for the forecast and gets three different answers depending on who is asked
You are paying for a GTM stack that nobody actually decided to buy
You are about to hire a VP of RevOps and you are not entirely sure what the job is
If four functions own revenue, nobody owns revenue
The first move isn't a system or a tool. It's a single owner with a real mandate from the CEO, sitting above sales, marketing, customer success and finance, and accountable for the whole arc from lead to renewal.
Your discount rate is the cheapest diagnostic in the business
Pull the average discount for last quarter. If nobody can produce it in an hour, you've found your governance gap, and it's costing you more than any lead-gen problem you're currently worrying about.
Retention you cannot see is retention you cannot manage
Before you hire another customer success manager, build the signal. A CS team without a health score is a team that apologises faster.