Reservation enables you to plan and buy your campaigns with a fixed cost, offering optimized reach and controlled ad frequency while helping you forecast your campaign's performance. This is similar to how people traditionally buy TV ads. This is a specialized, advanced option that most advertisers will only use if they want high assurance their ads reach a certain number of Accounts Center accounts.
Reservation works across ad types and devices. Since Meta targets based on real people, not cookies, we can more accurately predict reach and control frequency across devices.
CAN_USE_REACH_AND_FREQUENCY
parameter.stop_time
must be within 180 days of a prediction. GET
API call to https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>?fields=rf_spec
.target_spec
.user_os
, such as iOS_ver_2.0_and_above
.PacingType
field.Search rf_spec
for applicable limits:
Name | Description |
---|---|
array | Supported countries for reservation |
object | Minimum campaign duration in days, per supported country |
object | Maximum campaign duration in days, per supported country |
object | Days in advance campaign can finish at time prediction made, per supported country |
object | Minimum reach in number of Accounts Center accounts, per supported country |
Results look like this:
{ "rf_spec": { "min_reach_limits": { "US": 1000000, "CA": 1000000, }, "countries": [ "US", "CA", ], "min_campaign_duration": { "US": 3, "CA": 3, }, "max_campaign_duration": { "US": 30, "CA": 30, }, "max_days_to_finish": { "US": 56, "CA": 56, } } }
Predictions contain the number of Accounts Center accounts your ad can reach in a date range based on a given reach, frequency, audience and budget. Reservation estimates can help you simulate your campaign's lifetime results and will adjust based on your objective, budget, audience, formats and placements, brand safety, performance goal and frequency control settings.
Edits to a reservation campaign are possible after booking, but once the campaign is in flight, you cannot edit or pause it except to switch out your ad creatives. If you only edit an ad's creative, your prediction won't change. You may delete the campaign to cancel and stop it, but will need to re-book the campaign, where you may receive a new CPM and prediction on campaign outcomes. Only book campaigns you intend to run.** For testing, limit your reservations to the smallest size and length; be certain you cancel them since this is real ads inventory that we reserve for you.
These are default limits for predictions:
For details specify fields. To see all reachfrequencyprediction
s for an account, make a HTTP GET
to https://graph.facebook.com/{API_VERSION}/act_{AD_ACCOUNT_ID}/reachfrequencypredictions?fields={COMMA_SEPERATED_FIELD_LIST}
.
To reach all reachfrequencyprediction
s based on a reachfrequencyprediction
ID, make a HTTP GET
with fields you want: https://graph.facebook.com/{API_VERSION}/{RF_PREDICTION_ID}?fields={COMMA_SEPERATED_FIELD_LIST}
.
By default, Meta returns the ID. For field details, see Reservation Prediction, Reading.
This shows possible status
results in reachfrequencyprediction
. Initial limitations appear when applicable, however they may vary per ad account or by country in the future:
Code | Status | Description |
---|---|---|
1 | SUCCESS | Prediction successful |
2 | PENDING | Prediction still being produced |
3 | FAIL | Unreachable audience. Too high reach or budget. |
4 | FAIL | Prediction settings invalid, for example, duration |
5 | FAIL |
|
6 | FAIL | Budget or bid for given reach too low |
7 | FAIL | Too short ad set length |
8 | FAIL | Too long ad set length |
9 | FAIL | Ad set end date too far in future |
10 | FAIL | Frequency cap not specified |
11 | FAIL | Ad placement not supported, such as mixed RHS and Feed |
12 | FAIL | Ad set dates issues (start time and/or end time): Start time in past, not midnight, or not full day. End time in past, exceeds 90 days of start time or doesn't end after 6AM. |
13 | FAIL | Targeted country not yet supported |
14 | FAIL | Ad set dates include blackout days |
15 | FAIL | Insufficient inventory, unable to reserve. See Reserving a Prediction. |
16 | FAIL | Minimum reach required for account not achieved. See Getting Account Restrictions |
17 | FAIL | Actual reach available for this prediction is less than the minimum reach of the targeted country, usually 200,000 for most countries. |
18 | FAIL | Invalid day parting schedule provided. |
19 | FAIL | Target CPM unachievable. |
20 | FAIL | Frequency cap too low for blended delivery |
21 | FAIL | Ads inventory changed significantly enough for inaccurate prediction. |
23 | FAIL | Frequency cap interval not supported in target country. |
24 | FAIL | Holdout Lift Study ad set under account or campaign group not consistent with reservation prediction. |
25 | FAIL | Frequency cap can't exceed the number of days your campaign runs. |
26 | FAILURE_EMPTY_AUDIENCE | Selected audience empty and unusable. |
27 | FAIL | No modification allowed on your running campaign. |
28 | FAIL | Cannot modify running campaign created with Insertion Order. |
29 | FAIL | Cannot modify running campaign due to time constraints. |
30 | FAIL | To edit a running reservation ad set, choose a budget higher than current spend. |
31 | FAIL | Lift Study for account or campaign group starts after campaign starts. |
32 | FAIL | Lift study for account or campaign group ends before campaign ends. |
35 | FAIL | Cannot set Reservation campaign start time to be in the past. |
36 | FAIL | Please make sure the duration of the Reservation ad set is longer than one day and the campaign start/end time is valid. |
37 | FAIL | The objective isn't supported by Audience Network with the reservation buying type. |
39 | FAIL | Selected placements combination can't be used when buying with reservation. |
40 | FAIL | Specific mobile OS versions can't be targeted with the reservation buying type. |
41 | FAIL | Friends of connections can't be targeted with the reservation buying type. |
42 | FAIL | Reservation campaigns are not able to run when Audience Network is selected as the only placement. Please select the Audience Network placement with either Facebook Feed or Instagram Feed as additional placements. |
44 | FAIL | Reservation doesn't support Facebook Story. |
45 | FAIL | To use Facebook Stories as a placement, please also select either Facebook Feeds or Instagram Stories. |
50 | FAIL | Selected placements combination can't be used when buying with Reservation. For Reservation IO buying, please ensure the objective is Video Views. Otherwise, to use Facebook In-Stream, please select the Facebook Feeds placement. |
53 | FAIL | The in-stream video placement is available only for audiences in the US, the UK, Australia, New Zealand, Ireland, Thailand, Mexico, Peru, France, Germany, Argentina, Colombia, Spain, Chile, Ecuador, Dominican Republic, Guatemala, Bolivia, Honduras, El Salvador, Norway, Sweden, the Netherlands, Belgium, Poland, Portugal, Denmark, India, Malaysia, the Philippines, Indonesia, and Vietnam. To continue, edit your audience to include only people in those countries. |
60 | FAIL | To use Facebook Marketplace, please select the Facebook Feeds placement. |
66 | FAIL | Facebook Right Column Placement Cannot Be Combined with Other Placements. |
69 | FAIL | If you would like your ad shown on the Explore section of Instagram, you will also need to select Instagram Feed as a placement. |
100+ | FATAL | System failure, no user fault. Retry. |
Provide your Prediction ID and its data as input to create a new ID which serves as a reservation ID. Then attach this reservation ID to your ad set. Creating a reservation makes inventory unavailable to others, so you should attach it before it expires.
If the reservation succeeds, we temporarily reserve the inventory for you. You have approximately one hour after the reservation to assign an ad to an ad set.
If the hourly limit for reservation requests is exceeded, you'll see the following error code and message: 613: Calls to this API have exceeded the rate limit.
Reserve predictions for your ad sets to lock in your price and have predictable reach. Reserve an audience identified by reachfrequencyprediction
for a set time with reserve
for action
. You can use a single prediction ID to create multiple reservations. For example:
curl \ -F 'action=reserve' \ -F 'rf_prediction_id=<RF_PREDICTION_ID>' \ -F 'access_token=<ACCESS_TOKEN>' \ 'https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>/reachfrequencypredictions' // Response {"id":9876543210"}
To reserve inventory based on a prediction, make a POST
API call to https://graph.facebook.com/{API_VERSION}/act_{AD_ACCOUNT_ID}/reachfrequencypredictions
.
For reach
, budget
, and impression
, you can reserve a prediction at a specific point on curve_budget_reach
rather than the original tuple of prediction reach/budget. Use these fields:
Name | Description |
---|---|
int | Required.
|
string | Required for reservation and cancellation. Options are:
|
int | Optional. Reserved prediction or reservation ID. A new reservation releases a reserved audience and uses it for the new reservation. See Reusing Reserved Audiences |
int | Optional. ID of previously created prediction. New predictions use the audience from a given prediction. Note: |
int | Optional. If specified, provide Specify |
int | Optional. If specified, provide Specify |
int | Optional. If specified, provide You can override this value. To do so, specify the |
Meta reserves predictions asynchronously; you should poll and check the status of the prediction. Initially the prediction status is 2
(PENDING). On completion, status is 1
, SUCCESS
, or 15
, FAIL
which means we lack inventory to complete this reservation.
Since the reservation system is dynamic, you may see small changes in inventory availability between your prediction time and reservation time. However, Meta respects values you get at the prediction time, so long as changes fall into a reasonable threshold.
After you successfully reserve a prediction, create an ad set with it:
curl \ -F "rf_prediction_id=<RF_PREDICTION_ID>" \ -F "access_token=<ACCESS_TOKEN>" \ "https://graph.facebook.com/<API_VERSION>/<AD_SET_ID>"
To successfully assign a prediction, your ad set must meet these criteria:
start_time
- derived from predictionend_time
- derived from predictiontargeting
- derived from predictionbid_amount
optimization_goal
lifetime_budget
or daily_budget
rf_prediction_id
which attaches the prediction to the new ad set. buying_type
is RESERVED
You can also attach reachfrequencyprediction
to ad sets to modify its prediction. Making a POST
request to https://graph.facebook.com/{ad_set_id}
with rf_prediction_id
for the reachfrequencyprediction
you want to use.
The following are limits on the ad set:
facebook
, instagram
. feed
, and rightcolumn
. stream
, story
, explore
, explore_home
and reels
. If placement includes instagram
, you must use destination_ids
, not destination_id
. The destination_ids
field should contain the Facebook Page ID used as destination_id
, plus the Instagram account ID. promoted_object
must match the prediction's destination_id
. For page posts, it must be the page ID specified and for app ads, it must match the app ID specified.We charge reservation campaigns on actual impressions delivered. If the campaign start time passes and the ad set lacks active ads, the campaign fails to deliver and no charges apply. Meta releases remaining inventory, however we may penalize the ad account for repeat occurrences.
When you create a campaign using Meta's reservation buying type you are agreeing to pay the proposed costs for the advertising inventory you reserve. If you want to change your audience size or ad frequency, then your costs will also change. You can make these changes any time before your campaign starts. You can edit your ad creative until your campaign starts.
Reservation ad sets may contain multiple ads; you can add more ads at any point. If the ad set activates and there are no active ads in it, you must create your first ad within 24 hours for ad sets lasting 3 to 30 days, or 6 hours for ad sets lasting 1 to 2 days. If you do not, we delete the reservation.
You can make edits or pause your reservation campaign before it starts. Once the campaign has started, you can only edit your ad creative, your budget or extend the end date. If you only edit an ad's creative, your prediction won't change.
If you change your budget or end date, this may generate a new prediction for your campaign which you will see in reservation estimates.
You cannot pause your reservation campaign after it starts, but you can delete the campaign to cancel your campaign at any time.
To pause or edit a set after it starts, see Pausing or Restarting Running Ad Sets and Editing Running Ad Sets. To delete an active set, see Ad Set, Reference. You'll be charged for any impressions delivered.
Note, to avoid potential failures, we strongly discourage deleting all ads when a reservation ad set is live.
If you assign a reservation to an ad set before the ad set becomes active, you cannot change most attributes unless you detach the reservation. Make a HTTP POST
request to the set and set rf_prediction_id
to 0. You can only modify the name
attribute on the ad set object.
To detach a reservation:
curl \ -F "rf_prediction_id=0" \ -F "access_token=<ACCESS_TOKEN>" \ "https://graph.facebook.com/<API_VERSION>/<AD_SET_ID>"
Once an ad set is active, the reservation cannot be detached and attributes of the set cannot be modified, except a few ad attributes listed below that are on the allow list:
name
creative_id
creative_spec
conversion_specs
tracking_specs
view_tags
You can pause an active set, see Ad Set, Reference. If you pause an ad set for more than 30 minutes, we no longer guarantee the prediction for this set.
To reactivate a set paused for more than 30 minutes, you need a new prediction. Make a POST
to reachfrequencypredictions
. See Create a Prediction and Reserve the Prediction. You should pass an existing_campaign_id for the active set to be reactivated in the request. After you create and reserve a new prediction, attach reachfrequencyprediction
to the ad set with HTTP POST
to: https://graph.facebook.com/{ad_set_id}
specifying rf_prediction_id
for the prediction you want to use.
You can make these updates after a set starts.
You cannot edit or pause and active set if it meets one of the following criteria:
To edit running ad sets, get a new prediction. See Create a Prediction and Reserve the Prediction. You should pass an existing_campaign_id for the active set to be reactivated in the request.
After a new prediction is created and reserved, you can attach reachfrequencyprediction
to the ad set by making a HTTP POST request to: https://graph.facebook.com/{ad_set_id} specifying rf_prediction_id
as the id of the reachfrequencyprediction you want to use.
If you cancel a reservation it frees reserved inventory for other advertisers. However, you can reuse an audience from a previously reserved prediction if you aren't already using it. This enables us to take into account additional inventory to create a prediction, without you having to cancel an existing reservation.
Include rf_prediction_id_to_share
when you create a reservation. This is the ID of a previous prediction. This invalidates the previous reservation, so you can use this inventory for your newly created reservation.
To reserve the new prediction, you must also pass the additional parameter rf_prediction_id_to_release
which is the ID of the previous reservation.
You can rotate ads in the ad set you are using. You do not need to detach the reservation from the ad set to do this. Add one or more ads to the ad set and wait until it becomes active. At this point, you can change the status of the initial ad to paused. You must have at least one active ad within the ad set.
You can design a sequence of ads that deliver in order. First create the ad set and ads. Then specify the sequence at the ad set level in creative_sequence
. Each individual ad in this ad set may not appear, appear once, or appear multiple times in the sequence.
If creative_sequence
array length is zero, then you're using no sequencing. If the length is not zero, we recommend you make it equal frequency_cap
in rf_prediction_id
. If the length is larger than frequency_cap
, we truncate the last several ads from the array. If the length is less than frequency_cap
, we recursively auto-fill the array by repeating the sequence from the beginning. To make results clear, set the length of creative_sequence
to the same amount as frequency_cap
.
Each ad in the sequence has ACTIVE
, PENDING_REVIEW
, or CREDIT_CARD_NEEDED
status. A particular ad in the sequence will only be delivered to a user if all the preceding ads in the sequence have been delivered. Ads not included in the creative_sequence
will not be delivered.
All ads in an ad set using ad sequencing, no matter in the sequence or not, cannot be paused, archived, or deleted.
This feature is only available for reservation ad sets, i.e. the buying_type
of its parent ad campaign is RESERVED
, and this ad set has rf_prediction_id
set.
More details can be found at ad set document.
For predictable reach on Instagram, you can create a Reservation campaign with buying_type
set to RESERVED
.
Reservation helps you plan and book awareness and engagement campaigns, optimizing for reach, ad recall lift and ThruPlay.
Reach estimates in Ads Manager and the API can give partners guidance on what they can reasonably expect. The Instagram community comes first, we try to achieve reach objectives conservatively, and expect to evolve over time. All policies that apply to using reservation estimates for Facebook also apply to Instagram.
Code | Description |
---|---|
1487583 | An ad set with no ads cannot be assigned a reservation prediction |
1487055 | Ad set status invalid |
1487600 | Ad set is already assigned to a reservation. If you want to use another prediction, please first disconnect current from the set using null value and then assign a new prediction. |
1487578 | The specified |
1487581 | Reservation prediction cannot be modified on an active ad set |
1487594 | No ads in reservation ad set |
1487595 | Invalid Target Spec in reservation ad set |
1487614 | Ad set start time does not match with original prediction |
1487615 | Ad set stop time does not match with original prediction |
1487616 | Cannot associate ad set with invalid prediction |
1487671 | Direct transition from one prediction to another for an ad set is not allowed. |
1487244 | Ad set Update Failed - reason should be given in response |
1487672 | Failed to assign prediction to ad set. |
1487680 | You don't have permission to use reservation ad sets. |
Creating a reachfrequencyprediction
for an app destination_id
:
curl \ -F 'target_spec={"geo_locations": {"countries":["US"]}, "age_max":35, "age_min":26, "genders":[2], "publisher_platforms":["facebook"], "facebook_positions":["feed"]}' \ -F 'start_time=1388534400' \ -F 'end_time=1389312000' \ -F 'frequency_cap=4' \ -F 'reach=1000000' \ -F 'budget=3000000' \ -F 'destination_id=<APP_ID>' \ -F 'prediction_mode=1' \ -F "objective=MOBILE_APP_INSTALLS" \ -F 'access_token=<ACCESS_TOKEN>' \ 'https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>/reachfrequencypredictions' {"id":"67890123456"}
Creating a reachfrequencyprediction
for a page destination_id
:
curl \ -F 'target_spec={"geo_locations": {"countries":["US"]}, "age_max":35, "age_min":26, "genders":[2], "publisher_platforms":["facebook"], "facebook_positions":["feed"]}' \ -F 'start_time=1388534400' \ -F 'end_time=1389312000' \ -F 'frequency_cap=4' \ -F 'reach=1000000' \ -F 'budget=3000000' \ -F 'destination_id=<PAGE_ID>' \ -F 'prediction_mode=1' \ -F "objective=POST_ENGAGEMENT" \ -F 'access_token=<ACCESS_TOKEN>' \ 'https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>/reachfrequencypredictions' {"id":"67890123456"}
Creating a reachfrequencyprediction
for an app destination_id
with Instagram placement:
curl \ -F 'target_spec={"geo_locations": {"countries":["US"]}, "age_max":35, "age_min":26, "genders":[2], "publisher_platforms":["facebook","instagram"], "device_platforms":["mobile"]}' \ -F 'start_time=1388534400' \ -F 'end_time=1389312000' \ -F 'frequency_cap=4' \ -F 'reach=1000000' \ -F 'budget=3000000' \ -F 'destination_ids=[<APP_ID>,<INSTAGRAM_ACCOUNT_ID>]' \ -F 'prediction_mode=1' \ -F "objective=MOBILE_APP_INSTALLS" \ -F 'access_token=<ACCESS_TOKEN>' \ 'https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>/reachfrequencypredictions' {"id":"67890123456"}
Poll the following endpoint via HTTP GET requests to retrieve the status until it is something other than 2
:
https://graph.facebook.com/67890123456?fields=status
If the status is 1
(successful), then this can be attached to an ad set or reserved.
Reserving a prediction:
curl \ -F 'action=reserve' \ -F 'rf_prediction_id=<RF_PREDICTION_ID>' \ -F 'access_token=<ACCESS_TOKEN>' \ 'https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>/reachfrequencypredictions' {"id":"9876543210"}
Poll the following endpoint via HTTP GET requests to retrieve the status until it is something other than 2
:
https://graph.facebook.com/<API_VERSION>/<PREDICTION_ID>?fields=status
If the status is 1
(successful), then this can be attached to an ad set. Let's set up your campaign structure by creating a campaign, an ad set, a creative, an ad, and assigning the reservation to the ad set.
Create an ad campaign:
curl \ -F "name=Test" \ -F "buying_type=RESERVED" \ -F "status=ACTIVE" \ -F "objective=POST_ENGAGEMENT" \ -F "access_token=<ACCESS_TOKEN>" \ https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>/campaigns {"id":"1122334455"}
Create an ad set:
curl \ -F "name=TestReachSet" \ -F "status=1" \ -F "campaign_id=<CAMPAIGN_ID>" \ -F "rf_prediction_id=<RF_PREDICTION_ID>" \ -F "access_token=<ACCESS_TOKEN>" \ "https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>/adsets" {"id":"09876543"}
Generate an ad creative:
curl \ -F "name=sample creative" \ -F "type=1" \ -F "title=hello world" \ -F "body=hi i'm an ad" \ -F "link_url="https://www.facebook.com/" \ -F "image_hash=4aca812b4eadb72818a2c4124abd121a" \ -F "access_token=<ACCESS_TOKEN>" \ "https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>/adcreatives" {"id":"1323123123123"}
Create an ad:
// Create an ad curl \ -F "name=my ad" \ -F "adset_id=<AD_SET_ID>" \ -F "creative={'creative_id':<CREATIVE_ID>}" \ -F "access_token=<ACCESS_TOKEN>" \ "https://graph.facebook.com/<API_VERSION>/act_<AD_ACCOUNT_ID>/ads" {"id":"3213213123"}
Assigning a new prediction to the ad set:
curl \ -F "rf_prediction_id=<RF_PREDICTION_ID>" \ -F "access_token=<ACCESS_TOKEN>" \ "https://graph.facebook.com/<API_VERSION>/<AD_SET_ID>"