LivePlan · Financial Forecasting · 2025
When the forecast met reality
I led design for LivePlan’s 2025 forecasting overhaul, bringing real accounting actuals into the forecast so owners could plan against reality. The craft landed for the customers it was built for — but it didn’t move the metric we were chasing, and why is the more useful story.
The Challenge
A forecast that couldn’t see reality.
LivePlan helps small businesses build a financial forecast — but the forecast and their actual financials lived apart. A separate dashboard compared projections to bookkeeping, yet inside the forecast editor, where you actually change numbers, you could see neither this period’s actuals nor last year’s. You were planning blind, and the everyday loop of comparing and adjusting was a multi-step, multi-app process:
User Journey
First,
I want to see how I’m tracking against my forecast.
That information is in the…
The bet behind the work was a strategic one. Our most valuable, longest-tenured customers don’t write a plan and leave — they manage against it, comparing forecast to reality every month. The thinking went that if we made that iterative work dramatically easier, we’d retain more users over the long haul.
Discovery
Listening to the people who live in their forecast
We talked to the customers who do this work every month — established owners reconciling plan against actuals. We didn’t have to dig for the friction. The same questions came up, unprompted, again and again:
“Why can’t I see my actuals where I’m planning?”
Real performance lived on a separate dashboard. The moment they sat down to change the forecast, the numbers they needed were a tab away — so they memorized, switched back, and hoped they had it right.
“Wait — what is that number?”
An actuals figure would look off, and there was no way to interrogate it without leaving for QuickBooks, working out which accounts fed that line, and running a report just to see what actually happened.
“Let me organize this the way I think.”
A long, flat list didn’t match how owners hold their business in their heads. They’d been asking for grouping and roll-ups for years — structure that mirrored their mental model, not their chart of accounts.
Key Insight
A forecast you can’t compare to reality is just a guess.
The job wasn’t to add a comparison feature — it was to collapse the distance between where you see your performance and where you plan your future, so reviewing reality stops being a chore and becomes part of the flow.
What We Built
Four moves to close the gap
Actuals and prior years now sit inline in the forecast editor, recalculating your future projections from your real past performance — where the editor used to run blind to both.
What It Proved
It worked — for the people it was built for
Among the customers who’d connected their accounting, the work landed. They stopped leaving the forecast to check reality, investigated their numbers in a click, and adjusted their plans in place.
30%
Fewer trips to the Dashboard
Once actuals lived inside the forecast, people stopped jumping out to the standalone Dashboard to see how they were tracking.
8 in 10
Groups created, then populated
Despite a design with no drag-and-drop or multi-select, most users who started a group filled it — added items ~50× for every one removed. The model clicked.
10k+
QuickBooks round-trips avoided / month
Every transaction-panel open is a time a user inspected their bookkeeping detail without ever leaving LivePlan.
And yet…
Churn remained the same
We built the right tool. We just built it for a moment most of our core users were months if not years away from. Engagement shrank at every step down the funnel — and the deeper the feature, the smaller the audience.
Meeting our actual users where they are
Most users simply weren’t there yet. They came to LivePlan to write a business plan — iterative forecasting sat steps, sometimes years, further down their journey, with plenty that could stall them along the way.
So I mapped that journey as a funnel, to show leadership just how far most users were from the moment we’d been building for — and why it was time to pivot:
Hover over a segment to see Pains, Needs, and Desires
The Pivot
Moving up the funnel
When the churn numbers came back flat, the same evidence that predicted it pointed the way out. Instead of building deeper for the few who’d arrived, we moved up the journey — toward the 30-to-90-day window where most users actually decide whether to stay. The brief became simpler and harder: reduce burden, eliminate experience rot, meet basic expectations.
That redirect is what produced the work I’m proudest of at LivePlan — the business-plan redesign that cut early churn, and Idea Canvas, which meets people before they even have a plan. The forecasting work didn’t grow the pie. It taught us where the pie actually was.
Reflections
Output is not outcome
We shipped genuinely good tools, and the customers they were built for loved them — but none of it moved churn, because we were too far down the journey funnel to matter. The lesson that reshaped my judgment: craft quality and business impact are different questions, and you have to keep asking the second even when the first is going well.
Disagree, then commit — and bring evidence
It was a reasonable bet — that better forecasting would retain users long-term — but the customers and data I had pointed the other way: most users churn long before they have actuals to forecast against. I made that case early, committed fully, and helped build the best version of the bet we could. The lesson: disagree early with evidence, commit completely once the call is made, and keep that evidence close so the team can course-correct together when results come in.
Whose needs you optimize for is a design decision
Our loudest, highest-value users were a tiny, unrepresentative slice. Listening to them was right; treating their needs as the whole story was the trap. The fix wasn't to listen less — it was to map the needs against the journey, and notice that the leverage on retention sat at the top of the funnel, not the bottom.
A Vision for the Future
AI that helps you forecast
The loudest signal in our research wasn’t for another table — it was for help building the forecast in the first place. We’re now exploring an AI assistant that helps owners draft and refine a forecast without losing control of the numbers — the same craft thread, make the numbers approachable, applied earlier where more people can feel it.
