The company had closed a £4M Series A nine months earlier. ARR was just north of £2M. The team had eight account executives, a marketing manager, and a founder who was still personally involved in roughly 60% of closed-won revenue.

There was no RevOps function. No one owned pipeline governance, forecasting, or data quality. The CRM existed, technically, but it functioned as a contact database with an opportunity object bolted on as an afterthought. Pipeline stages were the Salesforce defaults. Close dates were fiction. The forecast was whatever the founder told the board it would be.

They didn’t have a revenue engine. They had a founder with a Rolodex and eight reps doing their best without a system to operate within.

I came in as a fractional RevOps operator. Two days a week, twelve months. The brief was simple: build the operating foundation that would make a Series B possible within 18 months.

Here’s how that year played out.

The Starting Point: Honest Assessment

Before building anything, I spent two weeks doing nothing but diagnosis. Sat in deal reviews. Read every CRM record from the last twelve months. Interviewed every rep, the marketing manager, and the three CS team members. Pulled the billing data and reconciled it against the CRM.

The findings were not unusual for a company at this stage, but they were severe.

  • Pipeline data was unusable. 40% of open opportunities had close dates in the past. Stage definitions were the generic defaults, and reps used them inconsistently. Two reps never updated stages at all.
  • No source attribution existed. Marketing was running campaigns but couldn’t connect spend to pipeline. Every opportunity was sourced as “inbound” regardless of whether it came from a webinar, a cold outbound sequence, or the founder’s LinkedIn.
  • The founder was the sales process. Pricing decisions happened verbally. Discounts were approved via Slack message. Deal terms varied wildly because the founder negotiated each one from scratch.
  • No forecast methodology. The board received a number each quarter. That number was the founder’s gut feeling calibrated by optimism. There was no pipeline coverage model, no weighted forecast, no commit/upside/best-case structure.
  • Comp plans were flat OTE with a basic quota. No accelerators, no segment differentiation, no SPIFs. Every rep had the same target regardless of territory quality or account mix.

The company wasn’t broken. It was operating the way every post-Series A company operates when nobody has deliberately designed the revenue architecture. The difference between companies that scale and companies that stall is whether someone builds that architecture before the next raise.

Months 1 to 3: Foundation

The instinct at this stage is to start buying tools and hiring. I refused. The first three months were entirely about design and cleanup.

Data model and pipeline architecture

I redesigned the CRM data model from the object level. Defined five pipeline stages based on the buyer’s actual decision process, not generic sales methodology. Each stage had entry criteria, required fields, and exit conditions. If the data wasn’t entered, the opportunity couldn’t progress.

This was unpopular for about three weeks. Then reps started realising that the structure made their lives easier because deal reviews became faster when everyone was speaking the same language.

Source attribution framework

I built a simple attribution model. First touch for pipeline credit, last touch for closed-won credit, with a multi-touch view available for analysis. No expensive attribution tool. Just disciplined UTM structures, form-level tracking, and a set of source values that marketing and sales agreed on.

The key decision: we would capture attribution from this point forward with clean data rather than try to reconstruct the past. Historical data debt would remain as a known gap. Four clean quarters would be enough for the raise.

Basic reporting

I built five reports. Pipeline by stage and rep. Pipeline created this month by source. Conversion rate by stage. Average deal size by segment. Forecast versus actual for the prior quarter (reconstructed manually as a baseline).

Nothing fancy. But the leadership team had never seen any of these numbers before. The founder’s reaction to the first pipeline review was telling: “I had no idea we had this many stale deals.”

Months 3 to 6: Architecture

With the foundation in place and clean data starting to accumulate, I moved to the structural work that would define how the revenue engine actually operated.

Compensation plan redesign

The flat comp plan was producing flat behaviour. Every rep chased the same deals regardless of strategic fit. I designed a new comp architecture with three changes:

  1. Segment-weighted quotas. Reps focused on mid-market had different targets than those working SMB. Quotas reflected the actual opportunity in each segment.
  2. Accelerators above 100%. The previous plan paid linearly. The new plan paid 1.5x above quota. This changed behaviour overnight. Top performers started pulling deals forward instead of sandbagging.
  3. Multi-variable crediting. New logo revenue and expansion revenue were credited separately. This aligned rep behaviour with the company’s need for both land and expand motions.

Forecast methodology

I implemented a three-tier forecast: commit, most likely, and upside. Each tier had rules. Commit meant a deal with a signed proposal or verbal agreement and a close date within the quarter. Most likely included deals in negotiation with identified budget. Upside was everything qualified but early.

The founder hated it initially. “I know which deals are going to close.” Perhaps. But the board needed a methodology they could evaluate, not a founder’s intuition.

Within two quarters, the forecast was accurate to within 8% of actual. That number would matter enormously during the raise.

Lead routing and handoff processes

Inbound leads had been going into a shared queue where the fastest rep won. This rewarded speed over fit and created a chaotic experience for prospects.

I built a round-robin with segment matching. Leads were scored based on company size and use case, then routed to the rep whose territory and expertise matched. Response time improved. Conversion from MQL to opportunity improved by 22%.

The sales-to-CS handoff was equally undefined. I designed a structured handoff process with a mandatory handoff meeting, a customer profile document populated from the CRM, and a 30-day post-close check-in owned by CS. Early churn, within the first 90 days, dropped from 11% to 4%.

Months 6 to 12: Scale Readiness

The architecture was working. Data was accumulating. The question shifted from “how do we operate?” to “how do we scale this and prove it to investors?”

Tech stack decisions

The existing stack was Salesforce (barely configured), Mailchimp (for all email), and a standalone dialler with no CRM integration. I made three deliberate changes.

Replaced Mailchimp with HubSpot Marketing Hub for marketing automation and lead scoring, integrated natively with Salesforce. Added Gong for call recording and deal intelligence. Kept the dialler but integrated it so activity data flowed into Salesforce automatically.

I did not add anything else. The temptation at this stage is to buy tools for every problem. Most of those tools create more data fragmentation than they solve. Three integrated platforms were enough.

First full-time RevOps hire

At month eight, the architecture was stable enough to hand operational ownership to a full-time person. I wrote the job spec, screened candidates, and onboarded the hire.

The critical insight on timing this hire: if I had recommended hiring a RevOps person at month one, they would have spent their first year building what I built in the first six months, but slower, because they wouldn’t have had the pattern recognition from doing this across multiple companies. The fractional model gave the company senior architecture at a fraction of the full-time cost, then transitioned to a full-time operator once the system was designed and running.

I stayed on one day a week for the final four months to mentor the hire and handle the investor-readiness workstream.

Territory model

With six months of clean segment data, I could finally build a territory model based on evidence rather than gut. I analysed win rates, deal sizes, and sales cycles by geography, company size, and vertical. Then I redistributed accounts across reps to balance opportunity more evenly.

Two reps had been sitting on territories with 3x the opportunity of their peers. Their “outperformance” was largely territory quality, not skill. Rebalancing created a fairer system and surfaced which reps were genuinely strong versus which had been riding a favourable book.

Data hygiene and investor readiness

The first 90 days had been about stopping the bleeding. The final quarter was about proving the patient was healthy. I built an investor data room with pre-built analyses: cohort retention by quarter, unit economics by segment, pipeline generation by source, sales cycle trends, and forecast accuracy over time.

Every number was sourced from the CRM. Every metric had a definition document. Every trend had at least four data points.

The Outcome

The Series B process started at month fourteen. It closed at month sixteen. £12M at a 15x ARR multiple.

The investor’s operating partner told the founder something I’ve heard variations of before: “This is the cleanest data room we’ve seen from a company your size.”

That sentence was worth millions in valuation. Not because the data itself was special, but because it signalled operational maturity that most companies at £3.5M ARR simply don’t have.

Here’s what the numbers looked like at twelve months versus the starting point:

  • ARR: £2.1M to £3.5M. Growth accelerated not because we added reps, but because existing reps became more effective within the architecture.
  • Forecast accuracy: Unmeasurable to within 8% of actual, for three consecutive quarters.
  • Sales cycle: Unknown to 47 days average, with clear segment-level variation that informed territory and pricing decisions.
  • Pipeline coverage: Unknown to 3.2x, consistently maintained through governed pipeline generation.
  • Early churn (sub-90 day): 11% to 4%.
  • Rep ramp time: Estimated 9+ months to 4 months, because new hires had a system to operate within.

The RevOps function at month twelve was one full-time analyst, a fractional operator on one day per week, and a set of systems and processes that ran without daily intervention.

What Made It Work

Three things, in order of importance.

The operating model came before the tools and the team. I did not start by buying software or writing job descriptions. I started by designing how the revenue engine should work, then built the systems to support that design, then hired the person to run those systems. Most companies do this in reverse and wonder why their tech stack is a mess and their RevOps hire is overwhelmed.

The fractional model provided senior architecture at the right cost. A full-time VP of RevOps at this stage would have cost £120k+ and would have spent half their time on operational work that didn’t require their seniority. A fractional operator at two days a week delivered the architectural thinking the company needed while costing roughly a third of a full-time senior hire.

The founder let go. This is the one that most companies get wrong. The founder accepted that their personal sales instincts, while excellent, couldn’t be the foundation for a scalable revenue engine. They let the process replace their gut. They trusted the forecast methodology even when it disagreed with their intuition. They stopped approving discounts via Slack and started using the governance framework.

Building RevOps from zero is not a tools project or a hiring project. It is a design project. The architecture comes first. Everything else is implementation.