How to Optimise DV360 Campaigns: A Practical Checklist
To optimise DV360 campaigns, first confirm one primary KPI and working conversion tracking. Then follow a routine: check pacing and delivery daily, review bid strategy, inventory, audiences, frequency, viewability and creative weekly, and assess reach and learnings monthly. Change one thing at a time and give automated bidding time to learn before judging it.
Optimising DV360 is less about clever tricks and more about a steady routine. The campaigns that perform best are usually the ones someone checks properly every day and adjusts calmly every week.
Use this checklist as that routine. It assumes the campaign is built and live. If it isn't yet, our guide on how to create a campaign in DV360 comes first.
What should you check before optimising anything?
Make sure you are optimising toward the right goal, with data you can trust.
- One primary KPI per insertion order. Cost per lead, cost per completed view, viewable CPM or reach. Two competing goals in one place lead to muddled decisions.
- Working conversion tracking. Floodlight activities firing on the right pages, or GA4 conversions flowing in. Run a test conversion before launch. Our guide to tracking conversions in DV360 explains the setup.
- A sensible structure. Separate line items for things you want to compare, such as prospecting and retargeting, or mobile and connected TV. Too many tiny line items split the budget so thin that none of them learns.
- Realistic timing. Google says automated bidding can take up to four weeks to learn and calibrate, and budget or setting changes can take up to 24 hours to apply.
The DV360 optimisation checklist
| When | What to check | What to do |
|---|---|---|
| Daily | Pacing against plan | Fix underspend or overspend at insertion order level |
| Daily | Delivery issues | Creative disapprovals, empty audience lists, deals not serving |
| Daily | Brand safety | Scan new sites and apps for anything unsuitable |
| Weekly | Cost per result by line item | Shift budget toward the stronger lines |
| Weekly | Sites, apps and exchanges | Exclude placements with poor results or odd traffic |
| Weekly | Frequency | Tighten caps where frequency climbs without results |
| Weekly | Viewability | Compare viewable rates by inventory source and format |
| Weekly | Creative | Pause weak versions and rotate in fresh ones |
| Monthly | Reach | Check unique reach and cost per person reached |
| Monthly | Audiences and deals | Retire weak segments and test one new idea |
| End of flight | Learnings | Write down what to repeat and what to drop |
The sections below explain the items that most often move results.
Want to see the numbers for your brand?
Get a free DV360 media plan from our team within one working day.
How do you fix pacing problems?
Pacing comes first because it decides whether the plan happens at all. DV360's Optimisation view flags insertion orders that are underpacing, so start there each morning.
Know your pacing setting. "Even" spreads spend across the flight but can fall short if inventory tightens near the end. "Ahead" spends slightly faster, up to 120% of the prorated target, to avoid underspend. "Daily ASAP" spends each day's budget as quickly as possible, which suits limited inventory but can use up a day's budget early in the day.
For underspend, loosen one thing at a time: widen an audience, add inventory sources, raise a bid or CPM cap, or lower a strict predicted viewability threshold. Google's own help suggests lowering the viewability threshold when a line item will not spend.
How do you choose and tune the bid strategy?
Match the strategy to the KPI. DV360's automated options include maximising a KPI while prioritising spending the full budget, maximising clicks or conversions while prioritising a target, and brand strategies that maximise viewable impressions or completed in-view and audible video. You can also cap the average CPM.
Google's guidance is that an aggressive target can cause underspend, and that maximising your KPI while prioritising full budget spend works best in most cases. So start there, watch it for two to four weeks, and add a target only once spend is healthy.
Avoid big swings. Google warns that drastic CPM or CPC changes on a line item during the learning phase can cause spend spikes. As a rule of thumb, change bids or targets by 10–20% at a time, and leave about a week between changes on the same line item.
How should you move budget between line items?
Move it on cost per result, and only once there is enough data to trust the gap. Here is an illustrative two-week snapshot from a lead campaign:
| Line item | Spend | Leads | Cost per lead |
|---|---|---|---|
| In-market audiences, display | ₹1,20,000 | 96 | ₹1,250 |
| Custom intent list, display | ₹1,20,000 | 60 | ₹2,000 |
| Retargeting, display | ₹60,000 | 80 | ₹750 |
| Total | ₹3,00,000 | 236 | ₹1,271 |
The tempting move is to pour everything into retargeting. Don't. Retargeting is capped by the size of its pool, and it depends on prospecting to keep that pool full. Double its budget and you mostly double its frequency.
A better move is to shift ₹40,000 from the custom intent line to the in-market line, keep retargeting at the level its audience can absorb, and rebuild the custom intent list with tighter keywords before testing it again. If the in-market line holds near ₹1,250 per lead at the higher spend over the next fortnight, move more. If its cost per lead climbs sharply, you have found the point where it stops scaling.
As a rule of thumb, wait for 20 to 30 conversions on a line before cutting it. Fewer than that and the difference may just be noise.
Which inventory should you cut?
Review results by site, app and exchange every week. Exclude placements that take budget with no results, show very low viewability, or have odd patterns, such as very high click rates with no conversions. Mobile apps with accidental clicks are a common culprit.
Keep exclusions in advertiser-level lists where you can, so every new campaign starts with them. Our guide to improving ad viewability goes deeper on the viewability side, and our guide to reducing CPM in programmatic advertising covers inventory choices that lower cost without hurting quality.
How do you optimise video and connected TV lines?
Judge video on what viewers actually watched, not on what you paid per thousand. For YouTube skippable ads, track cost per view and view rate. For OTT and connected TV, compare cost per completed view across apps and deals.
The arithmetic is simple: cost per completed view = CPM ÷ 1,000 ÷ completion rate. Take two illustrative lines. One costs ₹600 CPM with a 90% completion rate, which works out to about ₹0.67 per completed view. The other costs ₹350 CPM with 45% completion, about ₹0.78 per completed view. The cheaper CPM is the more expensive buy.
Two DV360 details matter here. By default, viewability targeting on a video line item is not applied to connected TV devices, and optimised targeting is not available on connected TV line items. So audience and inventory choices on CTV stay exactly as you set them, which makes your initial plan more important.
Are frequency and tired creative hurting results?
Often, and quietly. Watch frequency for each line item and for the campaign as a whole. If frequency climbs while the conversion rate falls, cap it lower or widen the audience.
For creative, compare results by version every week. Pause clear losers. On longer flights, plan a fresh version every three to four weeks, because the same audience stops noticing an ad it has seen many times.
What should you review at the end of each month?
Step back from the daily numbers and ask bigger questions. Did unique reach grow, or did you keep paying to reach the same people? Which audiences and deals earned their place? What would you stop doing next month? Write the answers down. They become the brief for the next plan. Our guide to creating DV360 campaign reports shows which reports to pull.
When we run campaigns for clients, the same numbers appear in our DV360 Reports Section, with daily delivery, budget against actual spend and breakdowns by creative, audience, geography and channel.
What mistakes should you avoid?
- Changing several settings on one line item at once, so you cannot tell what worked.
- Judging automated bidding in its first week.
- Optimising to CTR when the goal is leads or sales.
- Cutting prospecting because retargeting looks cheaper.
- Ignoring frequency until people start complaining about seeing the ad.
- Leaving optimised targeting on when the client approved a strict audience. Our guide to audience targeting in DV360 explains when to switch it off.
Next step
If your team is short of time for this routine, our DV360 Managed Service includes daily pacing checks and weekly optimisation by our specialists. If you would rather keep control and add an experienced pair of hands, a co-managed DV360 account splits the work between your team and ours.