Most growth roadmaps are inventories: everything the team intends to build this quarter, with no honest through-line connecting any of it to the numbers the company actually has to move.
I learned the alternative running growth planning inside one of the largest marketing-software companies in the world, Intuit Mailchimp, on the Web Marketing team. The web experience was the front door for millions of prospects, and everyone had an opinion about it. Paid media wanted new landing pages. Product wanted the homepage to tell their story. SEO wanted more content. Every team lobbed requests at the web team, and planning had become a queue of the loudest asks.
A queue of requests is not a plan. It is a reaction. The system I built to replace it turns a company's goals into a ranked quarter of experiments that the whole organization has already agreed to before anyone presents a slide. It works the same way in a four-person startup and a four-thousand-person company. Only the number of people you talk to changes.
Start from the company's goals
Every planning cycle I have seen go wrong started with the team's ideas. Every one that went well started with the company's actual input goals: the specific numbers leadership is accountable for this year. New bookings in a target segment. Conversion rate from a particular traffic source. International growth in named markets.
Those goals are the frame. They set the terms without being the plan themselves, and confusing the two is the first mistake teams make. Goals are the question the business is asking. The learning agenda is your ordered set of answers, each one testable.
Company goals are the inputs. The plan is the ranked list of experiments you believe will actually move them.
Interview the org before you plan it
Before I write a single line of roadmap, I interview every function that touches the funnel. In that engagement it meant sitting down with the paid-media lead, the SEO lead, the international lead, product managers, analytics, and the copy team. The same three questions each time: what is working, what is your best untested idea, and what do you wish the web team would stop or start doing.
The interviews do two jobs at once, and the second one matters more than it looks. The obvious job is research. The people closest to each channel already know where the leaks are, so their input produces better experiments than any plan I could write alone. The quiet job is ownership. People execute plans they helped build, and they resist plans that land on their desk fully formed.
A roadmap built in a vacuum is a wish list. A roadmap built from a room of stakeholders is a commitment.
I keep every contributor's input attributed as I synthesize it, so when a theme shows up across five interviews it carries the weight of five names rather than my opinion alone. By the end I have a distilled set of strategic learnings and a long list of candidate experiments, each traceable to the person and the data that raised it.
Ranking that list is its own discipline. I score each candidate the same way I sequence any fix: how much it could move a company goal, times how confident the evidence makes me, divided by the effort to run it. That math pushes the clever idea with thin support below the quiet idea that is nearly certain, and it gives every line on the agenda a defensible reason for its place in the order. When someone asks why their pet project is ranked seventh, the answer is arithmetic they can check themselves.
The workshop is where ownership gets locked in
The temptation is to take the synthesis and go write the roadmap. Instead I run one structured workshop with everyone in the room. Not a brainstorm. A brainstorm produces a wall of sticky notes and no decisions. This is a working session with a shared learning agenda going in, balanced breakout groups, and a one-page template each group fills out: the question, the hypothesis, the test, and the metric that would prove it.
By the end of the session, thirty people have argued their way to a shared set of priorities and committed to specific experiments in front of their peers. That public commitment is the point. It is far more durable than any buy-in I could manufacture later with a deck.
Present a learning agenda of questions
The last step is the one most teams get backwards. When it is time to present to leadership, the instinct is to show everything the team will build this quarter. That is an activity list, and activity lists are hard to fund because they promise outputs while staying silent on outcomes.
So I present a learning agenda instead: a short, ranked list of the questions we are going to answer, each with a hypothesis and an estimate of the revenue or conversion it could move. Same underlying work, completely different frame. One says, here is a pile of things we will make. The other says, here are the three things we are going to find out, and here is what each one is worth if we are right.
Executives do not fund activity. They fund learning: a ranked set of questions worth answering, each with a number attached.
The reframe also changes the room. A list of outputs invites debate about taste, whether the homepage should look like this or that, and those debates rarely end. A list of questions invites agreement, because almost nobody argues that a good question is a waste of time to answer.
This process is the planning layer of the wider system I described in the growth operating system. It sits downstream of the audit, which tells you where the business is leaking, and upstream of the weekly growth review, which keeps the experiments honest once they are running. The audit finds the problems. The learning agenda decides which ones you attack this quarter. The weekly review makes sure you actually learn from the attempt.
Run it yourself
You do not need a large company to use this. The steps are the same at any size:
- Write down the company's real goals for the period, the numbers leadership is actually accountable for.
- Interview everyone who touches the funnel. Ask what works, their best untested idea, and what to start or stop.
- Synthesize the input into ranked questions, each carrying a hypothesis and a rough estimate of what answering it is worth.
- Run one working session rather than a brainstorm, and have the room commit to the top experiments out loud.
- Present the ranked learning agenda to leadership, with a number next to each question, and let them fund knowledge.
The difference this makes is not subtle. A team running a learning agenda spends the quarter answering questions that compound into real understanding of the business. A team running an activity list spends the quarter shipping things and hoping. One of those teams gets smarter every quarter. The other just gets busier.
If your growth planning has turned into a queue of the loudest requests, this is the kind of system I install. Let's talk, or email me at karran@karrangupta.com.