The experience behind the work.

Cody Boyte is our principal. He has owned the marketing systems, led the people carrying quota, and built software used by the buyers on the other side of the call. Clients work with him directly.

Here are nine projects from that work. Each one shows the problem, what Cody did, and what changed.

Marketing

Axial · Head of Marketing Automation · 2011–2014

Marketing automation at Axial

A very small marketing team moved to a far more capable automation platform without stopping its campaigns.

The problem

Axial was a venture-backed startup with a very small marketing team. It needed to replace its old marketing automation platform with a much more powerful one and move quickly while both systems ran side by side.

What Cody did

Cody mapped the requirements, the data to move, the owners, the dependencies, and a plan to run both platforms at once. He rebuilt the funnel so that form fills, automated emails, lead routing, and a one-week follow-up for quiet accounts worked as one system.

He automated the work that repeated and fit a template: outreach for reports, events, webinars, and inbound leads. Replies from customers stayed with people.

What changed

The new platform went live two weeks after the contract was signed. The remaining flows were rebuilt over the next six weeks to keep the cost of paying for two platforms low.

Faster replies to leads and coordinated campaigns brought in more meetings and signups. Each salesperson went from hundreds of tasks a week to a handful of useful actions each day. Sellers hit their outbound targets and had more time for customer replies and live deals. Axial grew from $1 million to $20 million in ARR during his time there.

Axial · Head of Marketing Automation · 2011–2014

A live demand dashboard for sales at Axial

A monthly report that took two interns three weeks became a live dashboard that sellers used on calls.

The problem

Each month, two interns spent about three weeks building a company report from separate CSV files, Excel lookups, and pivot tables. Sales could talk about demand only through that monthly report. Nobody could point to a live or past opportunity during a customer conversation.

What Cody did

Cody loaded the incoming CSV files into a local database and added a way to query it. That alone cut the data build from three weeks to five hours.

He then turned it into a live internal dashboard that anyone in the company could query. Sellers could pull up specific live and past demand while they were on a call.

What changed

Monthly reporting went from three weeks to five hours. In the end, the demand data was available in seconds. Sales conversations moved from abstract reports to specific opportunities that a customer could take or had already missed.

The same data later became the base of a public project that brought in half of the company’s web traffic.

Jornaya · Head of Marketing · 2018–2019

An account-based marketing system at Jornaya

Shared account lists tied press, events, ads, and sales calls into one system that could be measured from first impression to revenue.

The problem

Jornaya was moving into the mortgage and insurance markets and needed each market to see it as the expert. Its marketing was a set of separate activities rather than a system. The website, the automation, and the data could not support coordinated targeting or reliable measurement of revenue.

What Cody did

Cody built one system that ran from authority to revenue. It drew on Jornaya’s own data, press, trade publications, speaking, targeted campaigns, and a flagship company event. He managed about $200,000 a year in paid media, handled a migration between HubSpot and Marketo, and set up multi-touch attribution in CaliberMind.

He replaced generic keyword campaigns with account-based programs built around press pushes, product launches, and conferences. Marketing and sales worked from the same account lists, messages, ads, and content for each named prospect.

What changed

Attribution showed that generic keyword and interest campaigns brought in leads that rarely closed. Campaigns aimed at named people and events brought in more new revenue even when they produced no lead at all.

Targeted accounts closed at higher rates after a sales call. Meetings booked at each conference rose from 1–3 to 15–20. At other events, meetings rose from 5–10 to about 20. The program produced five times as many press mentions and four times as many deals closed from events. It contributed to a $100 million-plus exit within two and a half years.

Jornaya · Head of Marketing · 2018–2019

Press and events at Jornaya

Jornaya’s own experts and data turned press coverage and a customer conference into a source of deals.

The problem

Jornaya was a small company selling into specialized markets such as mortgages, insurance, and cars. Its PR agency cost about $70,000 a year and had produced two or three placements. Journey Summit, the company’s conference, was run as a low-cost customer event.

The company needed to look like an established leader in its markets rather than a newcomer trying to get in.

What Cody did

Cody rebuilt PR around Jornaya’s own experts. He mapped the editorial calendars of the trade publications, matched their topics with Jornaya’s data, and put the CEO and specialists in mortgage, insurance, and automotive forward for quotes, articles, and regular columns.

He moved half of the annual marketing budget into Journey Summit. The event left a hotel ballroom for The Barnes, Seth Godin replaced the CEO as keynote speaker, and the program was rebuilt to draw the leaders of the market.

What changed

Within six months, press placements grew from two or three a year to two or three a month. The following year they reached 8 to 12 a month.

Journey Summit cost about $100,000 more and drew nearly every important company in the market. It helped sales close key targets. Deals sourced from events grew fourfold. One prospect closed quickly after meeting the author of an article he had just read.

Sales engineering

Oracle · Principal Sales Consultant and Team Lead · 2016–2018

Automated RFP responses at Oracle

A language tool found the likely answers to RFP questions so sales engineers could spend their time on the hard ones.

The problem

Sales engineers had to find and adapt technical answers for each RFP under tight deadlines. RFPs arrived in many formats that were hard to read reliably. A wrong answer would cost credibility with the buyer. The system had to tell trustworthy matches from weak ones.

What Cody did

Cody wrote a Python language-processing tool that returned the three most likely answers for each question and showed how confident it was in each one. People still wrote the final answers.

He converted every RFP into a single column of questions before it went into the tool. When the first command-line version saw little use, he rebuilt it as an internal web app. He kept 30 to 60 minutes of manual preparation in the process because it made the results more reliable.

What changed

RFP response time fell by 80%. In his own before-and-after test, a technical RFP went from six to eight hours of work to two to four. The tool answered 50% to 60% of the questions that closely matched past RFPs and saved four to six hours on each request.

The core team used it often and nearby teams picked it up. Nobody maintained it after he left. That limited how long it lasted.

Oracle · Director of Sales Engineering · 2019–2023

A shared deal knowledge base at Oracle

A secure, shared set of deal notes let a team of sales engineers cover for each other and coach each other.

The problem

Sales engineers on complex enterprise deals kept their preparation on their own laptops and shared decks and notes by email. Strict security rules blocked the usual file-sharing and tracking tools.

New people on a deal had no quick way to learn its history. Leaders could not easily step in on risky deals.

What Cody did

Cody built a secure knowledge base from Markdown files in Obsidian. Structured tables tracked accounts, opportunities, and deal progress. Each company had a folder for stakeholder notes, meeting history, client decks, work in progress, and next steps.

He trained the team and built habits around it: Monday prep, one-on-ones, and Friday deal reviews where sales engineers walked through live trackers, coached each other, and reused what worked.

What changed

With deal context always at hand, sales engineers could cover for each other and coach each other across more complex deals.

Within 18 months, the team went from the fewest deals closed and the lowest revenue to the most revenue of any group. Close rates rose and preparation got faster.

Oracle CrowdTwist · Head of Product · 2023–2025

AI-built demo environments at Oracle CrowdTwist

Working demos built with AI tools let buyers see the product in their own industry instead of imagining it.

The problem

Sales supported customers in more than a dozen industries. Enterprise evaluations relied on existing customer apps and on asking buyers to imagine how the product would work for them. The old demo environment could not show the workflows, brands, and customer experiences that a prospect needed to see.

What Cody did

Using AI coding tools, Cody set up a dedicated subdomain and built product-backed demo environments for six industries. He tried real integrations first. He then switched to simulated screens around the CrowdTwist touchpoints that mattered because they were fast to build and still credible.

He created shared components and documentation written for AI agents and taught sales engineers to rebuild sites with Codex. He also reviewed demos before they went live and helped with local builds for one-off cases.

What changed

Nearly every sales engineer used the environments. They helped win deals in four industries where the team had no strong demo before: restaurants, retail, hotels, and sports.

Prebuilt environments cut demo preparation time sharply. Sales engineers could focus on each customer’s use case and build on their own without waiting on engineering.

Product

Oracle CrowdTwist · Head of Product · 2023–2025

Rebuilding CrowdTwist around a real-time campaign engine

Putting a real-time campaign engine at the core of the product made loyalty campaigns quick to set up and kept the platform up through its biggest surges.

The problem

CrowdTwist sold its loyalty features as separate pieces: offers, activities, segments, purchase rules, and campaign execution. Customers had to set up each one on its own. Building a campaign took deep knowledge of the system.

Customer-specific campaigns ran as batch SQL written by consultants. Engineering kept tuning those runs as they grew more frequent and heavier. Some still took longer than the gap between cycles. Behind the scenes, each feature also had its own evaluations, triggers, and queues. All of it strained the platform as the customer base grew.

The product needed real-time experiences without disrupting existing work or turning each customer request into a custom build.

What Cody did

Over 18 months, Cody led product and engineering through a redesign of both the product and the system underneath it. He found a feature buried in the product that could serve as its core. He made it a system of its own and moved the rest of the product onto it.

He redefined real time as filtering a stream of events rather than running SQL faster. Filters on a single receipt or activity became settings in the campaign engine. They could combine items, subtotals, totals, and counts. Logic that spanned several events or several days stayed in SQL. Feature-specific queues gave way to one event-based pipeline.

Instead of rebuilding every capability, the team changed how the features linked together and how customers and staff structured campaigns. Cody brought product, engineering, consultants, and customers together on a migration path. New campaigns moved first. Existing campaigns that put heavy load on the system moved with them.

What changed

Most campaign updates went from one to four hours to a second or less. Setup went from three to seven days to a day or less. Sales demos became more flexible. Campaigns kept working when data arrived late.

One client doubled its program in about 36 hours and added millions of members and a surge of activity. Other systems failed under that load. CrowdTwist stayed up and processed the data without trouble. Holiday downtime and failures dropped sharply compared with earlier Black Fridays.

One customer replaced about 50 overlapping SQL queries with five campaigns and renewed several times. Another removed its custom platform code and the cost of maintaining it. A third launched dozens of campaigns it had promised its board in weeks instead of months. By the time Cody left, a multi-event version of the engine was scoped and ready for engineering.

Oracle CrowdTwist · Head of Product · 2023–2025

A roadmap customers could trust at Oracle CrowdTwist

Clear rules for the roadmap turned competing requests into a plan that customers and sales could trust.

The problem

CrowdTwist was a small team inside a Fortune 500 company. Requests from large clients, executive priorities, and promises made to prospects kept pushing planned work aside. Customers had lost trust in the product’s direction and saw development as slow and behind competitors. Some were leaving.

The product also offered several ways to reach the same outcome. That risked turning a product that could scale into a build-for-hire consultancy.

What Cody did

Cody set up stage gates with clear criteria for moving forward. He cut the product down to one clear path for each outcome and judged features by their cost and value to the whole market rather than by single requests. He tied market strategy to initiatives and features and rewrote the roadmap story to show how small releases added up to a larger plan.

Every two weeks, sales, customer success, and account management met to review requests, offer alternatives, and agree on how to describe them to customers. Monthly reviews kept decisions consistent and helped those teams set expectations they could meet.

What changed

Within a year, single-customer requests fell to about 10% of the roadmap as customers regained trust in delivery and direction. New deals rarely depended on promises of custom features. In the one case where a feature was a condition of signing, it moved up an item already on the roadmap instead of disrupting the plan.

In business and roadmap reviews, teams could show committed items or workable alternatives. Ad hoc requests, changes in direction, and stress on product managers all went down.