DSP optimization

The 6 hours your trader doesn't have

A programmatic trader's Monday is 8.5 hours long. Two of those hours are the work they were hired for. This is what the other six-and-a-half look like — with Outfox, and without.

A trader's Monday today

Without Outfox

4 hrs

2.5 hrs

2 hrs

Fire-fighting4 hrs
Meetings, reports, admin2.5 hrs
Actual optimization2 hrs

Monday · with Outfox

What Outfox clears off the trader's plate.

One Monday · 8 interventions

Fire-fighting
Weekend resets, pacing drift, broken caps — the things that quietly eat Monday morning.
4
fires
Reporting & admin
Reports, decks, broken templates — the second-most expensive hours of the week.
2
tasks
Real optimization
The work the trader was hired to do — with the right questions queued up.
2
plays

Fire-fighting

4 fires

Weekend resets, pacing drift, broken caps — the things that quietly eat Monday morning.

Sat · 11:42 PM· While the trader is asleepThe Trade DeskAnomaly

Geo target reset itself during a platform update

Outfox

Detected a silent geo reset on TTD line item LI-8841 (Aetherwave Q2 Awareness).

Targeting fell back to US-ALL at 23:42:11 UTC after a platform-side migration. Spend velocity has already jumped 312% vs baseline in the last 14 minutes.

Outfox

I paused the line item and snapshotted the previous targeting (NYC, LA, Chicago, Miami DMAs). Want me to restore it now or wait for Monday?

WinRestore targeting · log incident

Restored at 11:46 PM. $4,820 of misallocated spend prevented before Monday open.

Without OutfoxTrader discovers this Wednesday. Three full days of spend hitting the wrong DMAs.

Sun · 6:00 PM· End of weekendHeads up

Weekend pacing summary across all four DSPs

Outfox

Here's where the portfolio sits going into Monday.

✓ TTD — on pace (after Saturday's auto-fix)
⚠ Beeswax — underpacing 18% on Helix Athletic
⚠ DV360 — overpacing 22% on Northwind Insurance
◐ Xandr — reporting delayed 6h, eta 7:30 AM Monday

Outfox

I've drafted fixes for the two pacing issues. Approve them in the morning, or I can ship the low-risk one (Beeswax bid floor +12%) now and surface the rest for review.

WinShip Beeswax fix · queue the rest for 8 AM

One fix live before Monday opens. Two queued, ranked, and ready to approve in a single tap.

Without OutfoxTrader opens Monday morning and tabs through four dashboards to figure out what happened.

Mon · 8:47 AM· Trader logs inBeeswaxAuto-fix ready

One inbox, not four. One decision, not forty.

morning. what do I actually need to look at

Outfox

Three things, in priority order:

1. Helix Athletic — Beeswax bid floor draft is queued. Expected to recover pacing by EOD Tuesday. (Approve)
2. Northwind Insurance — DV360 overpacing is upstream of a frequency cap that didn't apply on a new audience segment. (Show me)
3. Aetherwave incident from Saturday — already resolved, post-mortem doc drafted in Notion. (Read)

WinApprove all · review #2 in detail

Cleared the morning queue in 4 minutes. Trader still has coffee in their hand.

Without OutfoxForty-five minutes of tab-switching, sticky notes, and 'wait, was this the one with the cap issue?'

Mon · 9:40 AM· Status call wrapsDV360Heads up

Northwind overpacing, explained in one screen

ok, the dv360 one. what's actually going on

Outfox

Frequency cap of 3/7d is set at the insertion order level. The new 'In-Market Auto Intenders' segment was added below it as a child line item — caps don't inherit downward in DV360. So that segment is uncapped.

Net effect: ~22% of impressions in the last 96 hours went to ~6% of users. Some have been served 28 times.

Outfox

Cheapest fix: clone the IO-level cap to the new line item. I've staged the change. One click and we're capped at 3/7d everywhere.

WinApply cap · regenerate reach forecast

Pacing back inside guardrails by 1:00 PM. Reach forecast emailed to the AM automatically.

Without OutfoxTwo hours of pulling logs, comparing line items, asking the AM what changed last week.

Fire-fighting handled.

Reporting & admin

2 tasks

Reports, decks, broken templates — the second-most expensive hours of the week.

Mon · 9:10 AM· Five minutes before the status callReport ready

Client report regenerated against the live campaign IDs

Outfox

The Aetherwave noon report template still references campaign 8741992 — archived March 14. Same template feeds Helix and Northwind reports.

Outfox

I've mapped each report to the current live campaign IDs and previewed the noon decks. The Aetherwave deck has one chart flag I'd want you to eyeball: CPM jumped Friday and the narrative slide still says 'stable.'

WinSend all three at 11:55 AM · ping me on the CPM slide

Three reports out on time. AM tells the client about the CPM bump before the client asks.

Without OutfoxTemplate silently pulls from an archived campaign ID. Trader notices on the call. Apologizes to the AM. Apologizes again to the client at 12:14 PM.

Mon · 12:55 PM· Lunch, properlyHeads up

New launch brief, prepped before the 1 PM planning meeting

Outfox

Heads up — 1 PM planning is the Aetherwave 'Summer Sustain' launch. I've drafted a starting point.

Pulled from the last three Aetherwave flights:
• Best-performing creative archetypes (top 3, with CPA)
• Suggested mix: 55% TTD CTV, 30% Beeswax PMP, 15% DV360 YouTube
• Two unknowns to discuss: incremental measurement partner and creative refresh cadence

WinOpen in the meeting deck

Meeting starts on slide 2, not slide 0. Ends 18 minutes early.

Without OutfoxTrader half-present in planning, still mentally untangling the morning's fires.

Reporting wrapped.

Real optimization

2 plays

The work the trader was hired to do — with the right questions queued up.

Mon · 3:00 PM· The first quiet stretchOptimization

Real optimization, with the right questions queued up

Outfox

You have a clear 2.5 hours. Three things are genuinely worth your judgment right now:

1. Helix CTV — completion rate dropped 14% last week on the 30s spot. Two hypotheses queued. (Investigate)
2. Northwind — the In-Market Auto Intenders segment is converting at 2.1x the legacy audience even after the cap. Want to shift 20% of budget? (Model it)
3. Aetherwave — a lookalike seed refresh is due. I've prepped three seed options ranked by expected lift. (Pick one)

let's do #2. model it at +20% and +35%

Outfox

Modeled. +20% holds CPA inside the client's $42 target with 88% confidence. +35% slips to $46 in 30% of simulations. Want me to stage +20% for tomorrow's daily review?

WinStage +20% shift · annotate the seed refresh thread

Two more decisions made before 5 PM. Trader closes the laptop at 5:18.

Without OutfoxTwo hours of optimization, mostly spent figuring out where to start.

Mon · 5:30 PM· End of dayOptimization

Tomorrow's first hour already drafted

Outfox

End-of-day recap is in your inbox. Two items waiting for you tomorrow:

• The +20% Northwind shift, ready to approve at 9 AM
• The Aetherwave seed refresh, three options with my pick ranked first

Nothing else is on fire. I'll wake you up if that changes.

WinSign off · let Outfox keep watch overnight

Trader leaves at 5:30 with tomorrow already half-written.

Without OutfoxTomorrow starts with checking pacing across four DSPs. Again.

Real work, finally.

The same Monday · with Outfox

From 2 hours of real work to 6.5.

Not because the trader is faster. Because the fires never started.

1.5 hrs

6.5 hrs

Fire-fighting0.5 hrs
Meetings, reports, admin1.5 hrs
Actual optimization6.5 hrs

−87%

Time spent fire-fighting

From 4 hrs to 30 min

−40%

Time spent in admin & reporting

Templates self-heal

3.25×

Real optimization time

From 2 hrs to 6.5

Your optimization works — we help with the rest.

The resets, the broken templates, the mystery numbers, the half-present meetings. Outfox is the second set of eyes that never logs off, on every DSP at once.

Book a 30-min intro →Talk to the team