2026 CampaignTech Awards · Engineering Team of the Year

Invisible
Complexity.

Four engineers built the infrastructure most campaign firms rent from a stack of outside vendors, and made all of it simple enough that the user never sees how hard it was.

AAAlex AndersonWLWalker LittleJOJacob OrellJBJessica Becks
co/efficientcoefficient.org
What they replaced

Most firms stitch together outside vendors.
This team built the whole stack in house.

Collection vendorData collection

More of it owned in house, so the team sees how the work is produced.

Dashboard vendorPolling engine

Intake, universe counts, lists and dashboards in one environment.

Modeling firmModeling platform

Scored audiences, live counts, message guidance and lists.

Texting platformco/nnect

Carrier registration, numbers and opt-outs, built from the ground up.

Advocacy toolAction Advocacy

Pages, trackers and client applications, generated in minutes.

Better, faster, cheaper, because we built it ourselves.
co/efficient2
Why this team

The hardest engineering is the part nobody sees.

A campaign operator gets three buttons. Underneath them, this team solved validation, compliance, cost, scale and failure, so the person sending the work never has to think about any of it.

co/efficient3
01 · co/nnect texting platform

What the operator sees: three steps.

Upload a list
→
Create the project
→
Send
What the engineering handles underneathMillions of texts a day
Before importShows how many contacts are sendable, opted out or invalid
VariablesCustom fields arrive with the list and filter inside the platform
TargetingFilter large universes, including on modeling scores, with no re-cutting
ContentConditional copy and creative by audience in a single project
PreviewRenders every personalized combination before it goes out
Segments and costCounts SMS segments and estimates cost, even as personalization changes length
ComplianceCarrier registration, number acquisition and opt-out management, built in
SafeguardsPauses on certain sending problems and alerts the team
co/efficient4
Hard problem · Reliability

When a send goes wrong, the system stops and asks a human.

On campaign deadlines, a bad send can't be undone. So co/nnect was built to catch certain problems itself rather than keep going.

1Detect

co/nnect spots certain sending problems as the campaign runs.

→
2Pause and alert

It pauses the campaign automatically and alerts the team in Slack.

→
3A human decides

Nothing more goes out until an operator decides how to continue.

co/efficient5
Hard problem · Architecture

Every message ever sent stays reachable, without paying to keep it all live.

The usual approachKeep every campaign in the active database.

Costs climb with every send, and the database gets heavier every cycle.

What Walker builtA data-lake architecture.

Historical message data stays accessible without sitting in expensive active storage. It's built for scale, lower cost and resilience as volume grows.

co/efficient6
02 · Polling engine

One environment, from final questionnaire to client dashboard.

  • Survey intake, universe counts, list workflows and dashboard production in one place.
  • Universe counts up front, so the team knows the exact scope before a survey goes out.
  • Surveys field the same day the questionnaire is final. Results in 48 hours.
A published release produced in the engine
A published release built in the engine: South Carolina's 1st District, October 2026.
co/efficient7
Hard problem · Editing at campaign speed

Researchers edit a finished dashboard by asking for the change.

An AI agent embedded in the engine takes plain-language requests to edit data and revise analysis. Editing is no longer manual, so results reach clients faster.

And with a recent engine update, several people can edit the same dashboard at once.

RequestShorten the labels
RequestRearrange the pages
RequestSplit one page into three
Kinds of edits the team makes by request. One request split a single page into three.
co/efficient8
03 · Modeling platform

The work campaigns send to a modeling firm, done in house.

  • Client-defined universes with live counts.
  • Message guidance for the exact audience selected.
  • Contact lists generated from the same environment.
Several weeks→ 2 business days
Every voter lands in a group, with a plan for each
Every voter lands in a group, with a plan for each. Illustrative data, fictional race.
co/efficient9
04 · Action Advocacy

An advocacy generator: from issue to message sent, built in minutes.

It creates pages, dashboards, live trackers and client-specific applications, with integrated data tools for awareness and voter ID.

1The issue, explained
2Choose email or text
3Pre-written and pre-addressed
4Sent
Demonstration page, fictional issue.
co/efficient10
What owning the stack delivers

Control over the code is control over speed, cost and quality.

Same daySurveys field the day the questionnaire is final
48 hoursFrom field to results in the client's hands
2 daysFor a voter model that used to take several weeks
MillionsOf texts a day through a platform we own
Faster

No waiting on a vendor's roadmap or another firm's queue.

Cheaper

No vendor markups or fees passed on to the client.

Better

Fewer handoffs, fewer errors, and full visibility into how the work is produced.

co/efficient11
Engineering Team of the Year

Four engineers. The scope of several vendors. None of the complexity reaches the user.

AAAlex AndersonWLWalker LittleJOJacob OrellJBJessica Becks
co/efficientInvisible Complexity