Open rate collapsed after a partner list was merged
Fabrikam Outdoors sends a monthly newsletter to 40,000 recipients from real-time marketing. Over three months the open rate fell from 34% to 19%, the complaint rate rose to 0.28%, and hard bounces sit at 5.8%. The change coincides with the addition of 25,000 addresses exported from a partner's webinar platform. The newsletter itself is a single 900-pixel-wide JPEG: headline, body copy and the call to action are all burned into the image. Outlook desktop users report a blank grey box, on dark-mode phones the logo sits on a white rectangle, and all links resolve through a tracking subdomain the team never set up or verified. Volume is flat at one send per month.
1. Rank the contributing causes and state the order in which you would fix them. Justify why content fixes should not come first.
Read the worked answer
Ranking by damage per day of delay
- List quality (the partner list). 5.8% hard bounce and 0.28% complaints are the two signals mailbox providers act on, and both are far outside safe thresholds. 25,000 addresses that never opted in to Fabrikam are the obvious source. This is the cause that compounds: every additional send to them raises the reputation penalty applied to the 40,000 who did opt in.
- Sending and link domain authentication. Links on an unverified tracking subdomain, plus whatever the SPF/DKIM/DMARC state is for the sending domain, affect inbox placement and therefore every metric downstream. Cheap to fix, permanent benefit.
- Content and rendering. Image-only email with a 900px hard width is genuinely broken in Outlook and hostile to dark mode, but it degrades engagement, not deliverability.
Why not content first
A beautifully rebuilt newsletter still lands in the junk folder if reputation and authentication are poor; you would then be unable to tell whether the rebuild worked. Fixing the list and the domain first gives you a clean measurement baseline, then the content change shows up as a real lift. Also note the open rate is the weakest of your three metrics - image blocking and privacy-driven pre-fetching both distort it, while bounces and complaints are hard numbers.
2. Specify how you would rebuild the newsletter in the email designer so it survives Outlook desktop, image blocking and dark mode.
Read the worked answer
Structural fixes
- Abandon the single JPEG. Use a responsive layout around 600px with real text blocks: heading block, text block, image block, button block. Target a text-to-image ratio well above 50% text.
- Headline and call to action become text and a styled button, never pixels. A recipient with images off must still understand the offer and be able to click.
- Every image gets meaningful alt text, and no image carries information that exists nowhere else.
- Set a preheader so the inbox preview shows the offer, not 'View this email in your browser'.
Dark mode
- Define logo, fonts and palette in the brand profile so the same assets are used everywhere; supply a logo variant with transparent or explicitly themed background instead of a white PNG on white.
- Avoid hard-coded white backgrounds on containers that hold light-coloured logos; avoid pure black text on transparent backgrounds.
- Test in the designer's preview across clients, and send live seeds to Outlook desktop, Gmail app in dark mode, and an iPhone.
Links
- Verify the domain used for tracked links before the next send, and use HTTPS throughout. Broken or unverified redirect domains are both a trust signal and a click-measurement problem.
Process fix
Add a pre-send checklist to the approval workflow: images-off render, dark mode render, plain-text version, all links resolved, seed test to at least three clients. The reason this newsletter shipped broken is that nobody owned that checklist.
3. The team wants personalised greetings ('Hi {firstname}') and a conditional block for customers versus prospects. What must be true before that ships, and what fallback rules do you set?
Read the worked answer
Preconditions
- The personalisation field must be populated for the audience you are actually mailing, not for the CRM average. Pull a count of blank or null first names in the target segment first.
- The condition must key off an attribute that is reliably present and has a defined 'unknown' state - a lifecycle or customer-type field maintained by the integration, not a free-text industry field.
Fallback rules to set
- Default value on every dynamic field: 'Hi there' rather than 'Hi ,'. Never let a templating miss produce a visible brace or an empty gap.
- Conditional block gets a third branch for unknown values, and the default branch must be the safe message, not the most aggressive one. A prospect shown a customer-only renewal offer is a worse outcome than a customer shown a generic offer.
- Same rule for the subject line: if the personalised subject has no fallback, skip personalising the subject at all.
Verification
Send a seed list that deliberately includes a contact with blank first name, a contact with no lifecycle value, and a contact whose email contains a plus-address. Review the rendered output before the live send. Personalisation failures are rare per recipient and therefore never show up in aggregate metrics - only in complaints and unsubscribes.