Download CV
Back to all case studies

Tax, accounting and advisory for business owners and nonprofits.

One lead pipeline and two legally separate websites for a US accounting firm.

Industry
Professional services
Duration
1 year
Website
whipplewood.com
Services (6)
WhippleWood's work in one picture: the tax and audit service pages on a phone, the lead chatbot built on Zapier, the lead routing flow into Karbon, the Denver homepage, Google clicks that kept climbing through the domain move, 4 of 12 new leads from AI search, and the nonprofit page, each marked with the platform it ran on

The services

  • AI workflow automation

    131qualified leads generated
    ~56%raw inquiries auto-filtered
    12AI systems built
    • Owned the lead-to-CRM pipeline end to end: enrichment, scoring, duplicate detection and entity-correct routing, across all four client touchpoints (chatbot, website form, intake form, R&D calculator)
    • Diagnosed a structural gap in WhippleWood's CRM data model and wrote the data-entry standard that fixed it, adopted company-wide by their own operations team

    TODO: add a picture of this work to serviceImages in this client's file.

  • AI assistants and chatbots

    140conversations handled since launch
    • Owned the chatbot build end to end: two AI chatbots for WhippleWood, one per legal entity, each with its own knowledge base and conversation logic
    • Wrote the chatbot's entire knowledge base: structured content built for machine retrieval, not human reading

    TODO: add a picture of this work to serviceImages in this client's file.

  • Marketing analytics and optimization

    1business day, at most, to first response
    15minutes saved per lead on qualification
    3.5xwin rate improvement
    • Owned the sales dashboard and lead-visibility system end to end: built it, set the speed-to-lead standard behind it, and got a new lead in front of the team within a day instead of sitting cold
    • Set the reporting cadence the team now works to, turning raw pipeline data into a plain-language read leadership can act on

    TODO: add a picture of this work to serviceImages in this client's file.

  • Content strategy and production

    2case studies published
    24social posts published
    • Owned WhippleWood's content strategy end to end: content pillars, a weekly cadence and a full monthly calendar across blog, social and case studies, built from scratch
    • Run the ongoing content engine as editor: 4 quality-gated articles a month through an AI-assisted production pipeline, every piece reviewed before it publishes
    • Built the firm's first case study program, now running on a standing quarterly cadence through a real-interview process
    • Direct WhippleWood's social output: 8 to 12 posts a month, planned around the content pillars

    TODO: add a picture of this work to serviceImages in this client's file.

  • SEO, AEO and GEO

    3.0xnon-branded search clicks, in a year
    72%average Google position improved
    99%site health score
    2.3xsearch clicks on the evergreen guides
    24%average Google position improved
    • Owned WhippleWood's technical SEO end to end: audited the site, rebuilt tracking from scratch after the prior agency's access was lost, and ran a quarterly audit to keep it healthy
    • Led the technical side of a legally required split into two sites, protecting search visibility through the move
    • Built the site's AEO/GEO layer and tied AI visibility to real leads by tracking AI search as a CRM source
    • Owned SEO strategy and reporting for the account, from weekly client updates to the CEO board decks
    • Rebuilt all 12 of WhippleWood's blog guides end to end: a new topic map, new outlines researched from what clients actually search, structure, FAQs, schema markup, titles, body copy and updated figures
    • Mapped the blog's topics into a clear hierarchy, so each guide had a clear place and linked to the right neighbors
    WhippleWood search results: clicks on the rebuilt guides grew 2.3x (221 to 499 a quarter) through the 30 December domain move, 93% of URLs indexed after the split into two legal entities, mobile PageSpeed from 46 to 89, and 4 of 12 new leads from AI search in 2 to 16 June 2026

  • Growth and product marketing

    46prospects targeted
    5emails in the sequence
    • Owned the cold outreach test end to end: built the 5-email sequence, got sign-off from the client's own leadership, and ran it to a clean, on-schedule close
    • Built a 5-email sequence for a franchise-prospect list, each email addressing a different objection, from credibility to a benchmark-report close
    • Closed the campaign out with an honest result instead of letting it drift, and recommended moving budget to warmer prospects

    TODO: add a picture of this work to serviceImages in this client's file.

The challenge

WhippleWood is a US tax, accounting and advisory firm serving business owners, nonprofits and franchise operators. It runs as two legally separate companies under shared ownership: one for tax and advisory work, one for audit and assurance.

When I joined, WhippleWood had no dedicated marketing person. The firm already ran on Karbon as its CRM, but had never standardized how it used it, and nobody could say where new business actually came from.

The website had a deeper problem than falling traffic. One site represented two companies the law requires to stay separate, a compliance risk, not only a branding one.

And there was no proof anywhere. WhippleWood had never published a case study, had barely any real presence beyond the site and blog, and once it got a chatbot, that chatbot had nothing accurate to answer from.

I joined the account as one of several people on it and became its sole owner about a year in. I set the standard for how the team used Karbon, centralized every lead into one path through it, rebuilt the site's technical foundation, split it in two along the legal line, and stood up its first content program, including pages written for the new chatbot to read.

In its first year, that system generated 131 qualified leads across three channels, a year of SEO work tripled non-branded search clicks and moved average Google position from 38.9 to 10.9, and the firm published its first case studies and built a real social presence from scratch.

Where it started

  1. A CRM with no standard

    Leads from calls, referrals, the website and later a chatbot were logged inconsistently or not at all, and most records had empty fields.

  2. A data-model gap underneath it

    Custom fields added to a contact never linked to the deal it belonged to, so real reporting meant exporting and matching records by hand.

  3. A damaged technical foundation

    A prior agency left lost tool access, broken indexing and schema, and falling search traffic nobody had diagnosed.

  4. Two companies, one website

    WhippleWood's tax and audit sides must stay legally and financially separate, but ran on one website that didn't reflect that, a compliance risk, not only a branding one.

  5. No proof, nothing for the chatbot to read

    WhippleWood had never published a case study, and once it got a chatbot, that chatbot had nothing accurate to answer from, just guesses.

WhippleWood's homepage, full screen, captured October 2026
One of the 12 rebuilt posts, Restaurant financial benchmarks, on WhippleWood's Insights

My approach

Before building anything, I audited how WhippleWood actually worked a lead, how its site actually performed in search, and what its new chatbot would need to answer a question correctly, three separate audits that kept turning up the same root cause: nothing had been built as a system. Karbon stored data nobody had standardized, the site carried years of neglected technical debt on top of a looming legal split, and there was no real content anywhere, for a person or for a bot. I treated all three as one project, because each depends on the others.

I fixed the lead system first: diagnosed a structural gap in Karbon's data model, wrote the standard that fixed it, then built the intake automation and the chatbots that file every lead correctly and route it to the right one of WhippleWood's two legal entities.

Alongside it, I rebuilt the site's search foundations, recovering a year of technical damage, then split the site in two along that same legal line without losing the search visibility I had just rebuilt.

And I built the content strategy that makes the other two work: a strategy and editorial calendar from scratch, the firm's first case studies, and pages written specifically for the chatbot to read, so it answers from real facts instead of guessing.

  1. Audited the lead process, the site's search performance and the chatbot's content gap, all before building anything
  2. Diagnosed the Contact-to-Work-Item gap blocking real reporting, then wrote the fix
  3. Built the chatbot and intake automation that files every lead correctly, routed to the right entity
  4. Recovered a year of technical SEO damage, then split the site in two along the legal line
  5. Built WhippleWood's content strategy and case study program from scratch
  6. Wrote the chatbot's knowledge base so it answers from real content, not guesses

My tool stack

I built WhippleWood's lead system on Zapier, Karbon and Slack, and measured the search work in Search Console, Google Analytics 4 and Semrush.

Research and keywords

  • Semrush

Crawling, indexing and speed

  • Google Search Console
  • Bing Webmaster Tools

Tracking and AI visibility

  • Google Analytics 4

Building and shipping

  • WordPress

Workflow automation

  • Zapier

CRM

  • Karbon

Email and outreach

  • Smartlead

Team approvals

  • Slack

AI models

  • ChatGPT (OpenAI)
  • Perplexity
  • Claude

Data and dashboards

  • Looker Studio

The impact

The clearest change is how little a new lead costs WhippleWood's team now. Someone used to have to notice an inquiry, decide if it was real, and type it into Karbon by hand; now that happens before anyone sees it, and the person who does see it is already looking at something qualified. Win rate on closed deals rose sharply after the system went live. That's a real-world before-and-after, not a controlled test: the pre-launch sample was small, and other things at the firm changed that year too, so the system doesn't get full credit for the lift.

The SEO story is about timing. The technical rebuild had to land, and the gains had to be real and dated, before a legally required domain split put a year of search visibility at risk, and the site came through that move stronger than it went in. What this doesn't prove is a direct line from search traffic to revenue: no organic-only lead count exists yet to connect a search click to a signed client, so the honest read is that the site got easier to find and faster to load, and the business case for what that's worth still has to come from the lead data, not the search data, alone.

The content program is the newest piece, and it shows in how it's measured. The firm now has real proof to send a prospect and a chatbot that answers from facts instead of guessing, neither of which existed a year ago. What it hasn't produced yet is engagement data: how many people actually read a case study, or how a social post performs, isn't tracked yet, and won't be until the account owner finishes compiling it later this year. That's a gap worth naming rather than covering with a vanity count.

  1. Every qualified lead filed in Karbon automatically, routed to the right company
  2. Win rate on closed deals improved 3.5x since the system went live
  3. Two legally separate websites, each with its own clean technical foundation
  4. The blog back above its pre-move traffic within three months of the domain change
  5. The firm's first case studies and its first real social presence
  6. A chatbot that answers from a knowledge base written for it, not guesses
Nancy, WhippleWood's AI assistant, live on the site and listing the firm's services
whipplewood.com on a phone, the tax, accounting and advisory company's site after the split, captured October 2026

Project results

  • Qualified leads generated, tracked by source

    December 2025 to October 2026, the system's first year. Source: Direct read of the #ww-prospects Slack channel, every lead classified by source.

    131

  • Of qualified leads from the website

    December 2025 to October 2026. Source: Direct read of the #ww-prospects Slack channel.

    113 (86%)

  • Of qualified leads from the intake form

    December 2025 to October 2026. Source: Direct read of the #ww-prospects Slack channel.

    12 (9%)

  • Of qualified leads from the chatbot

    December 2025 to October 2026. Source: Direct read of the #ww-prospects Slack channel.

    6 (5%)

  • Of raw website inquiries auto-filtered before reaching staff

    August to October 2026. Source: Direct read of #ww-prospects and #ww-prospects-unqualified Slack channels.

    ~56%

  • Of filtered-out inquiries that were vendor solicitations, not weak prospects

    August to October 2026. Source: Direct read of #ww-prospects-unqualified Slack channel, by rejection category.

    77%

  • Minutes saved per lead on manual qualification and entry

    Ongoing since the chatbot went live. Source: WhippleWood internal performance report, Jan 2026.

    15

  • Win rate on closed deals, before the system went live

    Deals started August 27 to November 19, 2025 (8 of 52 closed deals). Source: WhippleWood's own Karbon pipeline export (Google Sheet).

    15%

  • Win rate on closed deals, since the system went live

    Deals started December 2025 to October 2026 (162 of 300 closed deals). Source: WhippleWood's own Karbon pipeline export (Google Sheet).

    54%

  • AI-driven flows and systems built for WhippleWood

    October 2025 to October 2026. Source: Counted directly by the account owner, Oct 2026.

    12

  • Chatbot conversations handled since launch

    Roughly 10 months, since the chatbots went live. Source: Derived from the confirmed average of about 14 conversations a month, account owner, Oct 2026.

    140

  • Hours saved a month on manual qualification and CRM entry

    Average, December 2025 to October 2026. Source: Calculated from the qualified-leads total and time-saved-per-lead figures.

    3

  • Business day, at most, WhippleWood's speed-to-lead standard

    Ongoing. Source: Confirmed directly by the account owner, Oct 2026.

    1

  • Of the lead-to-CRM process automated

    Ongoing. Source: The account owner's own estimate, Oct 2026.

    ~90%

  • Non-branded search clicks, site-wide, in a year

    Jul to Sep 2025 against Jul to Sep 2026, same quarter year-over-year. Source: Google Search Console API, both domains combined.

    3.0x

  • Average Google search position improved, site-wide

    Jul to Sep 2025 against Jul to Sep 2026. Source: Google Search Console API, both domains combined.

    72%

  • Non-branded search impressions, site-wide

    Jul to Sep 2025 against Jul to Sep 2026. Source: Google Search Console API, both domains combined.

    +58%

  • Pages earning search impressions

    Jul to Sep 2025 against Jul to Sep 2026. Source: Google Search Console API, both domains combined.

    +74%

  • Mobile PageSpeed score

    Oct 2025 to Oct 2026. Source: Google PageSpeed Insights / Lighthouse.

    89

  • Site health score

    Oct 2025 to Jul 2026. Source: Semrush Site Health, confirmed in a weekly client update.

    99%

  • Keywords in Google's top 10

    Sep 2025 to Sep 2026. Source: Semrush keyword tracking, both domains combined.

    2.8x

  • AI citations a week

    Apr to Jun 2026. Source: A log-file analyzer the account owner built to track AI-crawler citations.

    2x

  • Non-branded keywords whose search result now shows an AI Overview

    Sep 2025 to Sep 2026. Source: Semrush keyword tracking, both domains combined.

    3.9x

  • Search impressions on nonprofit-focused pages

    The 30 days after a Nov 29, 2025 content and schema update. Source: WhippleWood internal performance report, Jan 2026.

    ~420%

  • Clicks on the same nonprofit-focused pages

    The 30 days after the same update. Source: WhippleWood internal performance report, Jan 2026.

    ~105%

  • Search clicks on the evergreen guides

    Sept to Nov 2025 against June to Aug 2026. Source: Google Search Console, both domains combined.

    2.3x

  • Search impressions on the evergreen guides

    Sept to Nov 2025 against June to Aug 2026. Source: Google Search Console, both domains combined.

    2.5x

  • Average Google position improved, on the evergreen guides

    Sept to Nov 2025 against June to Aug 2026. Source: Google Search Console, both domains combined.

    24%

  • Search clicks on the firm's strongest guide

    Sept to Nov 2025 against June to Aug 2026. Source: Google Search Console, both domains combined.

    2.9x

  • Case studies published

    December 2025 to October 2026. Source: Confirmed directly by the account owner, Oct 2026.

    2

  • Social posts published

    December 2025 to October 2026. Source: Confirmed directly by the account owner, Oct 2026.

    24