Marketing Cloud: Holiday Readiness Best Practices

Check your sending infrastructure
Simplify email complexity to improve build time and send rate
Optimize Journey Builder Performance
Optimize API performance and usage
Personalize your messages to improve Deliverability
Change management
Monitor your results
Trust site real-time Salesforce Status
Planning your timeline - sample

Check your sending infrastructure 

• Get ready
  Prepare messaging estimates for your peak marketing season. This information is used as an input into Deliverability planning, to support total volume goals. 

  • Prepare and sync your Marketing Plan for all channels (email, mobile, web, social, in-store).

• Domain(s) properly authenticated
  Any/all sending domains should be configured with SPF, DKIM and DMARC per Gmail & Yahoo’s email sender guidelines launched in 2023. Use the Private Domain product to add DKIM to your messages.

• Ensure you have enough IPs to support your peak sending volume

  • Shared IPs are for smaller senders only (Less than 100k per month).

  • If sending 100k or more emails per month, you'll benefit them to be on a dedicated IP.

  • If sending 250k or more email per month, Marketing Cloud requires you to be on a dedicated IP.

  • A single dedicated IP address can support ~2M messages/day once fully warmed. Plan ahead to have additional IPs warmed accordingly.
    

• Domain(s) and /IP(s) are warmed to volume requirements
  Your company’s sending reputation is tied to the domain(s) and IP(s) used to connect with your subscribers. Ensure that you are maintaining that reputation by not sending unexpected large volumes of mail which can look like spam to the MBPs. More on IP/domain warming can be found here.
  

• Start early
  If you need to add domains or IP capacity to your campaigns, plan to have them in place and warmed by early October. November and December tend to be the worst months to go live with a new IP address or new sending domain as mailbox providers (MBPs) are heavily scrutinizing traffic on their networks, making it harder to scale to higher volumes (i.e. warm an IP/domain).

• Other considerations for high volume
  Enhanced SFMC sending features such as Burst Sending or Dedicated MTA environments, or Catch and Release/Delayed Delivery that are ideal for some of our largest senders needing to send millions ++ of emails in a short period of time or avoiding sharing MTA bandwidth with other senders. Reach out to your AE for more information about these features.

Note: Marketing Cloud’s Deliverability Services is an optional offering that can provide dedicated monitoring, analysis and expertise of key deliverability KPIs, bounce and blocklists, inbox folder placement, and sender reputation. Reporting format, frequency and distribution is highly customizable, to fit any size client.  To find out more, please speak to your Salesforce Account Executive. 

Simplify email complexity to improve build time and send rate

• Optimize SQL queries
  When optimizing SQL queries, consider the amount of data you're working with, as well as the data model as a whole. Always try to join tables on a field that is a primary key as this is always guaranteed to have an index. A few small changes in query logic have the potential to drastically improve the execution time and ultimately progress to your send so it can go out as quickly as possible.

  • Use Querying for last mile segmentation.

  • WHERE clause is your friend.

  • Queries running for more than 10 mins are a red flag and should be evaluated.

  • Be mindful of the SQL query limits and guidelines.

• Tune your HTML and AMPScript 

  • Usage of Data Extension functions like InsertDE/UpsertDE function AMPscript is great, but if you have 3 or more of these calls in your code, you may notice your emails start building much slower. This is because all of these functions are queued up and executed at the end of rendering the email.  

  • Make the most of the RaiseError(). If you have bad data coming in, the RaiseError() function with a "TRUE" flag, allows you to fail a bad subscriber out of the send rather than failing the entire send based on potentially one bad record.

  • If you are writing send time values to a DE (such as email name, subscriber key, or order number), consider using send logging instead of something like InsertDE/UpsertDE, as send logging will more efficiently meet this requirement.

  • Move processing logic outside of the content block.

  • Use the new BuildRowsetFromJson AMPScript function as a faster and sometimes easier alternative to SSJS to parse JSON (Platform.Function.ParseJSON).

  • Test new content at scale prior to your highest volume event.  If you plan to send to millions+ of subscribers, this allows you to have a baseline for what performance is expected.  This also presents an opportunity to make tweaks to accommodate for performance challenges before prime time begins.

• Assess your data model

  • A Best Practice is knowing when to use deltas instead of passing in the full data sets: 

    • Deltas would be a daily file of only the new data for a single day, appending or upserting to a DE that already contains the last 29 days of data, rather than passing in a file with the last 30 days of data each time.

    • You would be sending a fraction of the data, which makes each step of the process faster and scales much better as well.

    • If a query copies or updates data from one DE to another in bulk, try creating an import definition in Contact Builder and selecting “From Existing Data Extension” as the source. Unlike queries, import activities are not subject to execution timeouts. This knowledge article covers a particular scenario for accomplishing this.

  • Apply retention to Data Extensions.

  • Avoid concurrent imports and utilize deltas instead of full data file refreshes.

  • Assess your data footprint with a Storage Report.

• Make sure images are only used when appropriate

  • Images should be no larger than 1200px wide at 72dpi.

  • Use images only for content that is a photo or an illustration.

  • Headlines, text, and buttons can be made of text and styled with code and will have very little size impact.

• Automation Efficiency

  • Stagger automation to avoid overlaps with your key campaign sends.

  • Minimize the number of queries and evaluate long-running automation that may no longer be needed.

Optimize Journey Builder Performance 

• Journeys to audit & optimize for holiday

  • Browse and cart abandonment journeys

  • First-time buyer journey

  • Post-purchase journey

  • Loyalty journey

  • Product returns journey

  • Transactional emails

• Optimize existing journeys and campaigns

  • Follow the Recommendations that Journey Builder provides when validating your journey. This helps avoid configurations that are known to impact scale and/or performance.

  • Create a pre-filtered, sendable copy of the data extension for each journey instead of using the same data extension for multiple journeys.

  • Incorporate Einstein Splits and Einstein Send Time Optimization in journeys.

  • Enable Journey Builder High-Throughput Sending (HTS) for journeys with email activities for better-sending performance and higher throughput.

  • Use system priority feature on the System Optimization dashboard to prioritize system resources for time-sensitive journeys during times with concurrent active journeys.

Optimize API performance and usage

• Identify which API best fits the use case.

  • Transactional Messaging API for transactional triggered sends vs Classic Triggers for promotional triggered sends

• Gracefully Handle Rate-limiting Responses.

  • Use back-offs and retries upon HTTP 429/503 (REST) & 500 (SOAP).

• Keep sustained rate for SOAP under 2000 req/min and REST under 2500 req/min. 

  • Build exponential backoff and retry logic in your clients to respond to rate-limiting errors.

• Batch API Calls

  • Replace singleton calls with batches up to 100 objects per call. 

  • Helps mitigate rate-limiting errors. For large DE reads and writes, use Imports and Extracts, instead of using singleton or batch reads and writes. Try to load as much data as possible into Marketing Cloud so you aren’t making a large rate of singleton external API calls.

• Use data import to move large or static content into the Marketing Cloud.

  • Use data extension update APIs only for small, dynamic data.

• Use synchronous API primarily for data retrieves, and other real-time needs. 

  • For example, for a real-time response for an action before the next action or a data retrieve

• Manage your DE size to minimize the latency of DE APIs. Stagger query activities and sends for best performance.

• Use asynchronous processing wherever possible.

  • Protects synchronous call performance for real-time needs like dynamic data retrieves.

  • Data writes, and triggers send calls.

    • For transactional triggered sends, and promotional triggered sends that need near real-time performance, use Transactional Messaging APIs.

    • For regular promotional triggered sends, use asynchronous SOAP or REST Classic Triggered Send API, or use Journey Builder entry APIs. 

For more information:

• API Best Practices and Use Cases

• Best Practices to Prevent Rate-Limiting | APIs | Marketing Cloud Engagement APIs and Programmatic Languages | Salesforce Developers

• SOAP API Performance | APIs | Marketing Cloud Engagement APIs and SDKs | Salesforce Developers

• Optimize API Calls and Data Structures to Improve Performance | APIs | Marketing Cloud Engagement APIs and SDKs | Salesforce Developers

• Client Timeout - SOAP, REST APIs
  

Personalize your messages to improve Deliverability

• Find your target audience
  Consider re-engagement campaigns for older list segments long before your peak season starts. Lapsed subscribers are more likely to mark your mail as spam, be an actual spam trap, or lead to block listing issues.

• Focus on engaged subscriber segments
  Engaged subscribers are typically considered your new sign-ups and those with an open/click/conversion in the last 6 months.

• Target content based on subscriber preferences and interests
  The more personalized the content, the better chance of converting a click to a purchase.

• Drive engagement for positive metrics
  MBPs can be quicker to block mail that has lower than typical engagement as they are on high alert for increased volumes and unwanted mail during this busy season.

Change management

• October, November, and December are the worst months to go live with a new IP address or domain name. Plan ahead for procurement, implementation, and warm-up lead times.

• Expect that many MBPs will be overwhelmed with excess volume sent by many senders during this time.  

• MBPs can be quicker to block, slow to respond, and a lot of sending has lower engagement. It's not the right timing for going through IP warming or redefining your marketing program.  

  • MBPs won’t issue exceptions during the holiday season.

  • MBPs postmaster contacts are not guaranteed, and only respond through proper channels and published processes, and some are closed for holidays and weekends.

  • Marketing Cloud does not have special access to MBPs, and cannot check your SPAM complaint rates for you.

• Level/ramp volume spikes as much as possible

  • General guidance is you don’t want to do more than double your daily send volume week-over-week. This means that if you send 20k/day today, you could jump up to 40k/day (max) immediately for a week, then 80k/day (max) the next week, etc. Larger and/or more frequent jumps can increase the likelihood of blocking issues. Sender reputation and list quality are the largest factors of just how quickly IP warming can occur.

  • Volume increases more than this will trigger MBPs automations, and potentially cause blocking issues.

Monitor your results

• What are your subscribers telling you
  The Email Performance Dashboard in Intelligence Reports is available to most Marketing Cloud editions and is one of the pre-configured dashboard options available to you. More information on how to get started with Intelligence Reports can be found here. Features like 'Drops in open rates' or 'clicks' may be your subscribers telling you they aren’t interested in the content being sent, listen to the feedback and adjust accordingly.

• Manage SPAM complaints
  Monitor complaints through mailbox providers feedback loops and reporting, and manage them, to keep them below 0.1%.  

• Notice a block or if open rates drop significantly
  Take a step back. What changed? Was an older subscriber segment used? Did the content or frequency change? Roll back any changes you may have made and review the guidance above, this isn’t the time of year for changing things up.

• Receive Send Error Notifications with Alert Manager
  Admin users can designate an email address to receive critical alerts for Email Studio, Automation Studio, and Journey Builder.

• Complete the picture of your Deliverability monitoring with feed based inbox monitors.

• Have a plan
  Understand that there is no 24x7 coverage at MBPs — they close for holidays and weekends like everyone else. Most MBPs utilize sophisticated algorithms that track your send history (IP/domain reputation) over the last 30 days and measure their user engagement (opens/clicks/complaints) with your email. The key to success in the inbox is sending highly relevant content to highly engaged subscribers, that’s not something an MBP will just take your word on. Salesforce does not have special access to MBPs to expedite unblocking. 

• Need some help?

  • Join the discussion in our Trailblazer Deliverability Group to stay up to date on the latest release details and ask questions or share ideas with your fellow marketers.

  • Didn’t make any obvious changes and you are sending to a highly engaged subscriber segment? Let’s take a look and see if we can engage the MBP on your behalf. Open a ticket with our Support team here, but do keep in mind, there likely isn’t a quick fix option. Consider suppressing the trouble domain or scaling back volumes while you continue your campaigns.

Trust site real-time Salesforce Status

• Transparent view into the status of the system

• Updates on system issues and incidents can be found here

• Automated system that facilitates an expedited view of incident

• Incident Notification Channel via Email

• Learn more on Trailhead

Planning your timeline - sample