How can we help?

Discovery's learning is primarily performed at the campaign level. Consolidating performance and data into a single account and campaign wherever possible enables smoother optimization. Therefore, operating across multiple accounts is not recommended.

We recommend using multiple campaigns when devices or creatives have different characteristics. Delivery may become concentrated on certain creatives within a single campaign. Splitting campaigns by messaging angle or device makes performance easier to compare.

PC and SP have different placements and user characteristics. Separating them into different campaigns makes operation, performance evaluation, and review management easier. If you want to compare performance by device, operate them as separate campaigns.

Geographic targeting can be configured at the prefecture level in Japan.

First, check the review status, daily budget, target CPA, placement restrictions, auction competition, and the campaign's learning status immediately after launch. New campaigns may take time to ramp up because of insufficient initial data.

A cold start is a state in which delivery does not increase easily because a new campaign or similar setup does not yet have sufficient data. It is more likely to occur when the target CPA is too low, the daily budget is insufficient, or the campaign is restricted by reviews or placement limitations.

If ad delivery is paused, it may take time for delivery volume to return to its previous level after the campaign is resumed. To maintain stable delivery, we recommend running campaigns continuously instead of repeatedly pausing and resuming them.

For tCPA bidding, set the daily budget at 5–10 times the target CPA. Even when the budget is limited, we recommend securing at least three times the target CPA or JPY 20,000 per day. Keeping the campaign from reaching its daily budget limit is more favorable for securing delivery opportunities.

As a general guideline, set the daily budget at 3–5 times the target CPA for MaxCV bidding and 5–10 times the target CPA for tCPA bidding.

If the daily budget is low, bidding may be restricted to prevent overspending, making it difficult to increase delivery. This reduces bidding opportunities, makes it harder to win placements that are likely to generate conversions, and increases the likelihood that the budget will not be fully spent.

When reducing the daily budget for tCPA bidding, limit each change to 20% or less and make adjustments no more than once every two to three days. Reducing it by 30% or more may cause delivery volume to drop. When increasing the budget, adjust it so that the campaign does not reach the revised daily budget limit on the same day.

The total budget is the maximum amount that can be spent across the entire campaign. Because tCPA delivery is optimized based on the target CPA, check the target CPA and daily budget in addition to the total budget when adjusting delivery volume.

When "Even Delivery" is selected, the daily budget is divided across the available delivery hours and an hourly budget limit is set automatically. Because an hourly limit is applied, bid adjustments and delivery opportunities are restricted. "Even Delivery" is not recommended for campaigns that prioritize conversions.

Set the target CPA based on the acquisition cost that you can realistically accept. If it is set substantially below actual performance or the market level, the campaign may struggle to win auctions and receive fewer impressions.

Limit each target CPA change to 20% or less and adjust it no more than once every two to three days. Frequent changes may require additional learning time, resulting in a higher CPA or lower delivery volume. Ideally, the target CPA should be close to the actual CPA. In particular, lowering it too far below the actual CPA may reduce impressions.

The target MCPA is not currently used for optimization, so it can remain unset at JPY 0. Updated in June 2026.

When using SmartBid, fixed CPC is not used for learning, so it can remain at the default value of JPY 25.

With tCPA bidding, the target CPA is a key setting that determines delivery volume. Check both the target CPA and the daily budget. Switching to MaxCV bidding may also be an option if you want to increase delivery.

A sudden increase in CPC is mainly caused by a temporary increase in auction competition for placements. If delivery becomes concentrated during highly competitive periods or on highly competitive placements, the likelihood of winning auctions increases, but impressions may also increase among users who are less likely to convert, resulting in a higher CPA. Possible measures include adding or improving creatives and dividing the campaign into two or three campaigns for A/B testing.

You can check the domains of websites where your ads were displayed in the management interface. However, you cannot view the specific ad slots where the ads appeared.

Ads will not necessarily appear on the same placements as in the past. Placements vary depending on each campaign's learning status and the ad slots expected to perform well at that time. Delivery on placements that previously performed well cannot be guaranteed.

You can configure a blacklist. Use the blocking function on the Sites screen to block specified domains. If you need to exclude a domain that is not displayed on the Sites screen, contact your account representative.

Excluding a domain before sufficient delivery data has been collected may lead to an incorrect decision. Before excluding it, check the amount spent on the domain, whether it generated conversions, its CPA, the current delivery stage, and performance in comparable advertising cases.

Whitelist availability depends on the bidding method. A whitelist can be configured for fixed CPC bidding. As a general rule, it cannot be configured for SmartBid. Contact your account representative if you would like to use one.

The recommended image sizes are 800×600, 600×600, and 300×250 pixels. These sizes are used by many accounts.

Video creatives are not supported. Animated images in GIF format can be used instead.

GIF files of up to 5 MB are supported. Because larger files may take longer to load and reduce CTR, we recommend keeping them at 2 MB or less.

An MSN ad may be rejected based on either Discovery's review criteria or MSN's review. First, check the review details displayed next to each asset at the ad level. Common reasons include a mismatch between the brand name and the PR disclosure on the advertorial landing page, missing company information on the advertorial landing page, or ad copy that does not meet the review criteria. Contact your account representative if the review details are unclear or the issue remains after revisions.

The MSN review begins after ad delivery is activated. The review may not have started if delivery has not yet been activated. Check the review status under "Review Details" in the management interface.

MSN may not disclose the specific reason for a rejection. In such cases, the cause cannot be attributed to a single factor. Review the ad copy, images, advertorial landing page, brand name, and PR disclosure based on past review patterns. As delivery continues and more data is collected, it may become easier to narrow down the cause.

Placements other than MSN use different review criteria and are generally less strict. Registering several creatives for review can help you understand the review tendencies.

For delivery through SSPs, the brand name in the management interface—the advertiser name entered during registration—must exactly match the PR disclosure on the advertorial landing page. A mismatch may result in rejection.

The main items reviewed are the creative, headline, and advertorial landing page. The brand name itself can be the advertiser name, product name, or a combination of both.

Types of ad copy that are frequently rejected in MSN reviews include definitive claims such as "always" or "guaranteed," medical or pharmaceutical claims, specific figures such as age, weight, or number of days, negative wording, and excessive use of symbols. We recommend avoiding these expressions or replacing them with more moderate wording.

Write MSN headlines at approximately 25–35 Japanese characters so that the entire headline can be displayed. Excessive use of symbols such as "!?" or "【】" may result in rejection.

Users currently cannot directly check whether an image has been rejected. Check the review status under "Review Details" in the management interface, and contact your account representative if the status is unclear.

The ad may not pass the review if the landing page does not contain company information or if the reviewer cannot access it. Check whether the advertorial landing page includes company information or a link to a guidelines page or a page related to Japan's Act on Specified Commercial Transactions. Also ensure that the landing page can be accessed from overseas IP addresses.

If an English-language review status or error is displayed, the creative or landing page, including its domain, may be affecting the review. First, check the review details shown next to the ad. Contact your account representative if the information is unclear.

A yellow exclamation mark indicates that the ad has been blocked on some placements or SSPs. Because the detailed status cannot be checked in the management interface, contact your account representative if necessary.

First, check whether sufficient spend has accumulated and whether delivery is concentrated on creatives or domains that have not generated conversions. Then consider pausing or excluding those creatives or domains and relaunching them in a new campaign.

CPA may fluctuate over short periods on specific placements such as MSN. Do not make a decision based on a single day's performance. Review performance over an appropriate period, such as one week to one month. Then reassess underperforming creatives and placements and consider improving the creatives or advertorial landing page.

To increase delivery, consider raising the daily budget, reviewing the target CPA, or adding campaigns. Testing creatives with a different approach from the existing creatives in a new campaign makes performance easier to compare.

Check the images, ad copy, GIF creatives, and performance by placement, and compare multiple variations. Images attract attention, while ad copy manages users' expectations about the post-click experience. Testing GIF creatives is one option for improving CTR.

As delivery progresses, it may become concentrated on certain creatives. Pausing a creative with a high delivery volume will reduce the campaign's overall delivery volume, so check the impact before pausing it. If you have new creatives to test, you can also run them in a new campaign.

Check the tag installation location, the conditions under which the tag fires, the postback configuration, and the results of a test conversion. If the conversion tag is installed on both the advertorial landing page and the product landing page, conversions may be counted twice, resulting in abnormal figures.

When changing an API integration or tracking parameter such as a UTM parameter, perform a conversion test before starting delivery to confirm that conversions are measured correctly.

You can measure events other than CV and MCV by creating and installing an "Other Event Pixel" in the management interface. These events are used only for measurement and are not used to optimize ad delivery.