Nobody agrees on the numbers

Wild Hive runs on a financial picture that's spread across at least three places — Workamajig, Leanne's separate Excel workbook, and a handful of competing per-account budget trackers — and the numbers don't reconcile cleanly. Kim and Leanne ran the same Workamajig report and got different results because of access differences inside the tool. Leanne has stopped trusting Workamajig's own actuals-vs-billed view because "it never matches" and she doesn't know why, so she keys numbers into her Excel by hand. A $59,000 over-budget project surfaced only in one account manager's tracker — it wasn't visible anywhere else in the stack. The same visibility problem extends to the agency's biggest variable cost: contractor spend on writing and design averages ~$25K/month, but the writing-vs-design split can't be pulled without recoding the books, and contractor hours don't land in the system until invoices arrive at month-end.

"I would love for these numbers and these numbers to talk to each other, because the person who is managing this could then really, truly see... it looks like everything's okay, but it's not."Kim · Meeting 3

"Everybody relies on their budget tracker to tell a story, and that's where the problem is."Kim · Meeting 3

Every project starts from scratch

There is no codified "Wild Hive Way" of doing things, and the operational consequences show up everywhere — three monthly activity reports compared in Meeting 1 used three different fonts and three different structures, Amber's running answer to process questions is "it depends," and the proposal process veers from bloated (5-6 people, 65-70 hours) to threadbare (one person carrying 25 hours) with no reliable middle. Allison is now actively pushing to standardize the foundation while keeping the surface bespoke — but the standardization itself doesn't exist yet.

"We don't have, like, a Wild Hive way. We have certainly some really beautiful slides that we'll use in a new business pitch, and we say, like, this is the Wild Hive way for strategic planning, or whatever, and we're just following the basic steps, but not... I can't really say that there's anything very codified over here."Allison · Meeting 2

"Everybody is not sitting around making activity reports with different fonts! Which just about made my head spin."Allison · Meeting 2

Scope creep quietly eats profit

When a client asks for more, Wild Hive often absorbs the extra rather than re-pricing it — and the loss only becomes visible after the fact, usually in one account manager's tracker. The single retainer-model client is the agency's biggest and least profitable account because scope shifts accumulated without budget retrofitting. The Raspberries 2024 over-budget event ("we sent too many people, nobody was caving") set an attitude that carried into other accounts. Kim is now pushing back on a specific account manager who took a project $55,000 over budget and blamed it on client demands.

"We need to discuss why you went over $55,000 on this project... 'the client kept wanting things'... that's not a client problem, that's an us problem."Kim · Meeting 3

"It's our least profitable account. It's our biggest and least profitable account, because scope shifts happened, and budget was not retrofitted to adjust to it."Kim · Meeting 3

Reporting and recap work consume hours that shouldn't have to be spent

The same shape repeats across at least four workflows: structured data exists somewhere, but a human is the assembly mechanism every cycle. Quarterly committee deck assembly involves "digging through 100 different files and folders." The weekly Prunes coverage email requires jumping between MuckRack, a bespoke spreadsheet, and an email composition step. The Potatoes USA scorecard requires Kim to manually count citations and people-talked-to every quarter, with a per-message scoring system on top of that launching next year. Amber tracks real-time metrics in "a good old spreadsheet" that's "all very manual."

"I have to count every person I talked to. How many citations of research that we touched in that quarter?"Kim · Meeting 1 (on the Potatoes USA scorecard)

"If there was a way for us to centralize all the, like, results, recap, photos, like, that kind of stuff, so that when the time does come to actually report on it, we're not, like, digging through 100 different files and folders."Kary · Meeting 1

The new-business pipeline is unpredictable

Wild Hive's proposal effort is "either very bloated, with a lot of people working on it, or one person kind of carrying it" — there's no reliable Goldilocks middle. Two recent pitches illustrate it cleanly: one took 65-70 hours across 5-6 people on a $200K opportunity, the other took 25 hours from Kim plus a hand-picked contractor on a $750K opportunity. The CRM is a spreadsheet that was last updated in December — at least 4-5 proposals from the past six months are missing from it, including two that are currently "still in the hunt." There's no qualification rule for what counts as a lead vs. a trade-show conversation, and win rate isn't computed.

"I feel like that's our proposal process all the time. It's either very bloated, and we have, like, a lot of people working on it, and we're all kind of getting us to the finish line, or it's, like, one person kind of carrying it and getting us to the finish line, in sort of a tight time frame."Kim · Meeting 3

Leadership can't see what's happening with clients or the team

Three visibility gaps surface from leadership directly. On clients: Allison openly asked "is that client happy?" with no answer, and a recent client-unhappiness event went on "way too long" before her quarterly call surfaced it. On team wellbeing: Allison doesn't feel she has "her finger on the pulse" of how individuals are doing, and the only bridge today is informal — Amber acting as Kim's mole into the junior staff. Kim has also named that several long-running clients have been flat year-over-year because Wild Hive doesn't always know what else they could be doing — they're "in nutrition jail" with some accounts that can't see the broader capability.

"I don't feel like I've got enough visibility or, like, really my finger on the pulse of, like, as much as I would like to have it. It's just... how's the team doing? How are our individual team members doing?"Allison · Meeting 2

"Amber is my direct report. Amber is very plugged into the junior staff... she's almost my mole, in a way."Kim · Meeting 2

Pick one financial source of truth — and clean up the coding

Right now three parallel financial pictures (Workamajig actuals, Leanne's separate Excel workbook, the per-account trackers) run in parallel and don't reconcile cleanly. Before any unified budget tool can be built on top of this, the team has to make a decision about which numbers are authoritative for which questions, and then enforce it. Inside that, two cleanup items are non-negotiable: stop using "mark as billed" on full-month actuals when the account is over budget — write off the difference instead — and recode the books so contractor spend can be split between writing and design going forward. Without these, any dashboard or budget tracker we build will inherit the same trust problems Leanne and Kim live with today.

Settle who owns what

A surprising amount of Wild Hive's operational work runs on informal ownership — Kim does the budget reconciliation because she cares about it, Emma was supposed to own the CRM but it wasn't explicit until Meeting 3 surfaced the mismatch, nobody owns "client health" today, and team-wellbeing signal travels through Amber as Kim's "mole" because there's no formal channel. Before tools can be built around any of these, leadership has to assign a named owner (and a documented backup) for each major area — client relationship health, BD pipeline cadence, scope-creep escalation, reporting standards, contractor management. Anything we automate will inherit whoever the owner is; if there isn't one, the tool drifts.

Write down the Wild Hive Way — at least the foundational pieces

Allison has already made the case that the foundation needs to be standardized even if the surface stays bespoke. The work is now writing it down — at minimum, the strategic-planning process, the scope-bucket → status doc → activity report linkage, the reporting baseline package (what every client gets vs. what's a priced add-on), and the proposal process. Some of this exists in people's heads (Kary articulated the scope-bucket framework cleanly in Meeting 1) but not in documents the team works from. The brand-voice IP, USDA-compliant nutrition methodology, and three-tier contractor model are similar — real institutional knowledge that lives in heads. Anything we build to "do the Wild Hive Way" is just an empty pipe until the methodology is written down.

Define what counts as a lead — and set a real CRM cadence

The CRM spreadsheet is woefully out of date because nobody agreed on what belonged in it. Kim raised the question directly in Meeting 3: are trade-show conversations leads, or are they conversations? Before any pipeline-health view can be built, the team needs to settle what qualifies something to enter the tracker, how often Emma updates it, what triggers a stage change, and when something moves from stalled to lost. None of this is hard — it's a 30-minute meeting between Kim, Emma, and Allison — but without it, any pipeline dashboard will show whatever shape the spreadsheet happens to be in that week.

Build the scope-creep muscle

The pattern Kim is pushing back against — "the client kept wanting things" becoming a $55K overage — is a discipline gap, not a tooling gap. Allison has already named the shape of the fix: when a client asks for biweekly status calls on top of monthly, that becomes "here's what your new program management budget will look like." But the team has to actually start saying it, with language ready, an approval path defined, and a tracker that flags scope creep when actuals start outrunning budget mid-engagement. The retainer-model account that's currently the biggest and least profitable is the proof-point — by the time the overage was visible, it was too late to reprice.

Centralize where the raw material lives

A meaningful share of the reporting-burden problem comes down to the assembly step — quarterly committee decks involve "digging through 100 different files and folders," media coverage requires jumping between MuckRack and a bespoke spreadsheet, photos and headlines live in different places across SharePoint and Google Drive. Before we can build a committee-deck assembler or a coverage-email drafter, Wild Hive has to make decisions about where assets, photos, results, scorecards, and message-hits live for each client. Not a single all-clients-everywhere solution — per-client is fine — but a documented per-client convention. Otherwise we'd be building a tool that has to guess where everything is.

A unified financial picture

A custom Workamajig connector plus an Artifact dashboard that runs Kim's mid-month pacing routine automatically, plus a weekly scheduled agent that surfaces accounts running hot. One connector, one dashboard, one agent — the build does what Kim does manually today, on schedule, with leadership visibility.

The Wild Hive Way, made executable

One flagship skill — the strategic-planning skill Allison named directly in Meeting 2 — plus one supporting skill that links scope buckets to activity reports. Authored in working sessions with Allison and Kim, not in advance of them. Each skill ships when its underlying methodology decision is settled.

Reporting and recap automation

Two skills that hit the most concrete time-sinks named. The Potatoes USA scorecard skill handles the manual counting today and the per-message scoring system launching next year. The MuckRack coverage email skill operates on a manually pasted MuckRack export — no new platform connector required for v1.

BD pipeline visibility

A skill that operates on the existing CRM spreadsheet — enforces the qualification rules from Move 04, helps Emma keep it current, surfaces win rate. Plus a scheduled agent that nudges when stages haven't moved and an Artifact dashboard view of pipeline state. The spreadsheet stays. We make it do its job.

Content production and brand voice

The Writing System scoped to Potatoes USA as the pilot client only — voice profile drawn from the existing brand book, anti-AI rules calibrated for the client (GFF's em-dash callout becomes a hard-banned token in the GFF variant later), mandatory post-draft audit. Plus one per-client brand skill driving the existing HTML Deck and HTML Asset engines for Potatoes USA deliverables.

Design capacity, augmented

The templated asset side only for v1 — decks, long-form pages, and ad-hoc branded assets for Potatoes USA, riding the per-client brand skill from Build 05. Establishes the per-client design pattern that everything else hangs off.

Client and team health surfacing

A scheduled agent that watches account-level signal across the data we'll already be pulling for other builds — Workamajig actuals (Build 01), pipeline movement (Build 04), and any other connected sources — and surfaces likely-at-risk and likely-upsell accounts to Allison and Kim on a defined cadence.