How we helped an engineering firm add ~$1.98M in annual retail margin.

Faster drawings still left them with the same margin per store.

Their retail team designed stores for national chains. A design could be adapted across many locations, and the firm earned a fixed fee for the work. Saving hours on each store should have improved the margin on the job.

Their in-house automation did save time preparing the first drawings. But retailers changed designs during rollouts, and their team absorbed work that hadn't been priced as an amendment. Permit comments sent drawings back for another round, while sets waited for a small group of licensed engineers to review and seal them.

As a result, later work consumed the early gain. Their leaders had to account for the technology budget without a clear answer about project margin. Their retail sponsor saw an over-budget store at closeout, when the extra hours had already been spent. The tools were in use, but the cost of finishing a store hadn't improved.

We measured what happened after the time savings.

To find where the gain disappeared, we ran a three-week diagnosis on one live retail program. We interviewed delivery leaders, project managers, their quality lead, and a sealing engineer, then checked their accounts against the work itself.

Team members logged what they worked on and what they were waiting for. We checked those entries against timesheets and traced retailer requests through emails and meeting notes to see which became paid amendments.

The live program showed the current work, but we also needed to examine completed jobs. So we reviewed 18 months of finished retail projects, comparing planned and earned fees, hours after the first drawings went out, revision rounds, and permit resubmittals. We also checked use of their existing automation and the time spent reviewing its output.

That study located the loss: unpriced changes, repeat comments, and drawings reaching scarce licensed reviewers before they were ready. Their design tools already saved time. And so we recommended keeping those tools and building the missing steps around the work that consumed the gain.

We gave managers a chance to bill changes before the firm absorbed the work.

We connected their project and time records, revision and permit logs, and the inbox where retailer changes arrived. Those records let us bring the change requests and review history into the work their team was doing.

Unpriced changes needed a commercial decision, so we built a step that flagged retailer changes in emails and meeting notes and drafted a change notice. Project managers could see the request and decide whether to bill it while the work was still visible. The billing decision stayed with their project manager.

Repeat comments needed attention before another review round. Our second step checked a drawing set against the city's past permit comments and their own quality checklist before it reached a sealing engineer. We reused their existing automation wherever it fit.

Their senior engineers supplied examples of correct decisions, and we tested our additions on unseen examples against the diagnosis baseline. We also documented rules for AI use on client work. Their licensed engineers continued to review and seal the drawings.

We kept the design tools that saved time and added checks where that time was being lost: client changes and repeated work after the first drawings went out.

Managers could act before an overrun became a closeout report.

The program team used both additions on live stores, and we reported use and results weekly. But better checks alone wouldn't give project managers an earlier view of their budgets. We also helped them review the pilot stores' budgets each week, while they could still act.

One project manager had discovered an overrun in a monthly report too late to change it. With the weekly view, that manager acted before closeout on a later job. Their retail sponsor could also see unpriced work early enough for a manager to address it. The information reached the team while it could still affect the job.

Less unpaid work and more billed changes improved the margin.

The pilot cut unpaid rework by 33%, covering revisions after the first drawing issue, repeat review comments, and permit resubmittals. Their team spent fewer hours repeating that work, and the changes rolled out across retail programs.

The firm reported $1.984M, or ~$1.98M a year, in retail margin improvement. Work they'd previously absorbed now produced $1.2M a year in priced and billed changes. The other $784K was an estimated annual benefit from less rework, calculated from the pilot reduction using blended labor cost and applied across retail after rollout.

In the following quarter, their firm also collected payment 5 days sooner on average, weighted by invoice value. They waited less time for money they were owed. Based on their observed receivables, that meant $1.1M in cash arriving earlier. That collection improvement changed when cash arrived and sits separately from the annual margin figure.

Their leaders could now connect the technology work to project economics. At a later technology-budget review, they had the retail result and its source page to show what had changed.

We gave their team a method they could use beyond retail.

We transferred the tools and the measurement method to their developers and research team. Alongside maintaining the tools, they had a way to examine a whole job, locate the loss, and decide what needed to change.

Their digital owner had faced requests for more tools without proof that the existing investment improved retail margin. After handover, their own team used the same discovery method on mission-critical work without us leading it. They could take what we'd provided into another part of the business.

See how we tested the checks and measured the results.

The technical companion explains the work study, drawing checks, human decisions, annual margin calculation, and cash-collection measure.

The time saved in design finally improved the finished job.

Their firm had already invested in automation that worked, but later work kept consuming its value. We found where that happened and built the missing steps around their existing tools.

As a result, their retail pilot needed less unpaid rework, their firm billed changes it had absorbed, and managers could act before closeout. The reported margin improved, cash arrived sooner, and their team kept both the tools and a method they'd already used on another workflow. Faster design now translated into a better result for the business.