One Two Three Send

Author: onetwothreeadmin

AI writing prompt chains: when to split one request into three
Photo by Jackson Simmer on Unsplash

Most solo operators treat AI writing tools like search engines: type a request, hit enter, hope for the best. When the output is vague or generic, they blame the model or rewrite the prompt with more adjectives.
The actual problem is structural. You asked one prompt to do three jobs—research a topic, adopt a voice, and format output—and the model optimized for speed, not depth.
Prompt chaining splits a single complex request into a sequence of smaller, focused prompts. Each step produces an output that becomes context for the next. It takes longer to set up, but the quality gap is measurable.
When a single prompt isn’t enough
If your request includes the word “and” more than twice, you’re asking too much. A prompt like “Write a blog post about email deliverability and make it conversational and include three examples and format it with subheadings” forces the model to juggle competing priorities.
AI models don’t multitask well. They process tokens sequentially. When you load a prompt with multiple instructions, the model allocates attention unevenly. Formatting often wins over substance. You get clean HTML wrapped around shallow ideas.
Prompt chains work better for:
- Long-form content (800+ words) where structure matters
- Technical topics that need accurate detail before stylistic polish
- Repurposing existing content into a new format or voice
- Iterative edits where you want control over what changes
If you’re generating a tweet or a subject line, a single prompt is fine. For anything that represents your expertise to an audience, chain it.
How to structure a three-prompt chain
Start with research and structure. Your first prompt should ignore voice and formatting entirely. Ask the model to outline key points, list examples, or extract the core argument from source material you provide.
Example first prompt: “List eight specific reasons email deliverability degrades over time for solo operators. Focus on technical causes, not general advice. No introduction.”
The output will be dry and mechanical. That’s correct. You’re building the skeleton.
Second prompt: expand and refine. Take the list from step one, paste it into a new prompt, and ask the model to develop each point with specifics. This is where you add constraints like word count, example requirements, or technical depth.
Example: “Take this list and expand each point into 2–3 sentences. Include one concrete example or number per point. Write for someone who manages their own email infrastructure.”
Third prompt: apply voice and format. Paste the expanded draft and ask for stylistic changes, structural tweaks, or HTML formatting. Keep the edits narrow—if you ask for voice and reorganization and new examples, you’re back to a multi-job prompt.
Example: “Rewrite this in a direct, operator-to-operator voice. Use H2 subheadings for each of the eight points. Keep all examples and numbers intact.”
Each step produces a tangible artifact you can evaluate before moving forward. If step one misses the mark, you catch it before spending tokens on polish.
The handoff is where quality breaks
Prompt chains fail when you don’t carry enough context forward. If your second prompt just says “expand this,” the model has no memory of why you wanted those eight points or who the audience is.
Always restate key constraints in every prompt. Audience, purpose, and scope should appear in each step, even if they feel redundant. Claude and GPT-4 handle long context windows well, but they still weight recent tokens more heavily. If your formatting request is three prompts deep, remind the model what the content is for.
Copy-paste the output from the previous step directly into the next prompt. Don’t summarize it or assume the model will infer continuity. The chain only works if each link sees exactly what the prior step produced.
If you’re using Claude, the Projects feature can store your chain structure as reusable templates. Set up a project with your three-prompt sequence, and each new piece of content follows the same quality path without rewriting instructions from scratch.
When to skip chaining and use a single prompt
Chaining adds friction. If you’re drafting something disposable—internal notes, a rough outline for your own use, a placeholder headline—don’t bother. Single prompts are faster and good enough for low-stakes work.
Chaining also doesn’t fix a bad brief. If you don’t know what you want in step one, splitting the request into three steps just produces three mediocre outputs instead of one. Do the thinking before you write the first prompt.
For most operators, the inflection point is around 500 words and one hour of expected reader attention. Below that, single prompts are fine. Above it, chain.
If you want to see how other solo operators are structuring their AI workflows—and what’s working in practice—subscribe to One Two Three Send. Every issue covers one specific tool, tactic, or operational decision for people running content businesses.

Heads up — some links in this article are affiliate links. If you sign up through them, we may earn a small commission at no extra cost to you. We only recommend tools we use ourselves.
2 August 2026
SEO content briefs: when keyword clusters hide topic gaps

Most SEO content briefs start the same way: export a keyword cluster from Ahrefs or Semrush, group by search intent, then write to the highest-volume terms. The assumption is that if you cover the keywords, you cover the topic.

That works until you publish, rank on page two, and wonder why engagement stays flat. The problem isn’t your on-page SEO—it’s that keyword clusters surface search behavior, not reader need. And the gap between the two costs you traffic you should own.

What keyword clustering actually shows

Keyword research tools group queries by semantic similarity and SERP overlap. If “WordPress caching plugin” and “best cache plugin WordPress” return similar results, they land in the same cluster. The tool assumes they represent one topic.

But clustering collapses nuance. A solo operator searching “WordPress caching plugin” might want a feature comparison. Someone typing “best cache plugin WordPress” might need a recommendation for shared hosting with 512 MB RAM. Same cluster, different jobs to be done.

When you write to the cluster instead of the underlying question, you produce content that ranks for some queries and satisfies none of them completely. Google notices. Your bounce rate climbs. Your rankings stall.

Where topic gaps hide

Three places keyword tools consistently miss context:

Question variants that don’t share keywords. “Why does my site slow down after enabling caching?” and “WordPress cache plugin makes site slower” describe the same problem, but most tools won’t group them. If your brief only covers the positive case—how caching improves speed—you miss readers troubleshooting the opposite.

Implementation steps buried in forums. High-volume keywords like “WordPress cache setup” get clustered. But the specific friction points—”cache plugin breaks checkout page,” “how to exclude URLs from cache”—live in Reddit threads and support tickets, not keyword reports. If you don’t surface those preemptively, your guide stays shallow.

Comparison queries across categories. Someone evaluating caching might also search “CDN vs caching plugin” or “caching plugin vs managed WordPress host.” Keyword tools treat those as separate topics. But your reader is solving one problem: site speed. If your content doesn’t acknowledge the adjacent decision, you send them to a competitor’s comparison post.

How to audit a brief for topic gaps

Before you write, cross-reference your keyword cluster against three sources:

People Also Ask boxes for your primary keyword. Google surfaces questions that didn’t make it into your cluster. If five PAA questions address plugin conflicts and your brief doesn’t mention conflicts, add a section.

Support forum threads for the tools you’re covering. Search the plugin’s own support forum or Reddit for your topic. Sort by recent activity. The questions that appear repeatedly in the last 90 days are the ones keyword tools undercount—because searchers phrase them a dozen different ways.

Competitor content that outranks your target keyword. Open the top three organic results. Skim their H2s. If two of them cover a subtopic your brief omits—say, “how to clear cache automatically”—that’s a signal. Either the SERP is shifting, or your cluster missed a question searchers expect answered.

This audit adds 20 minutes to your brief process. It routinely surfaces two or three subtopics that change whether a post becomes a bookmark or a bounce.

What this looks like in practice

A client brief for “email service provider for small business” included 18 keywords, all variations on provider names, pricing, and features. The cluster topped out at 1,200 monthly searches.

The PAA box surfaced “can I use Gmail as an email service provider” and “difference between transactional and marketing email.” Neither phrase appeared in the keyword export. Both questions represented decision points that determined whether a reader would even evaluate the tools in the article.

We added two sections: one explaining why Gmail’s sending limits break at scale, one defining transactional vs. marketing email with named examples—Postmark for transactional, MailerLite for marketing. The post ranked #3 within six weeks. Average time on page ran 4:20, well above the site median.

The keyword cluster gave us the semantic frame. The topic gaps gave us the utility.

The brief is a hypothesis, not a spec

Keyword research tells you what people searched yesterday. It doesn’t tell you what they needed, what confused them, or what adjacent question determined whether they acted.

If your content process stops at clustering, you’re writing to the average of past behavior. The posts that rank and convert are the ones that close the gap between the search and the need.

Run the audit. Add the missing sections. You’ll know it worked when your dwell time climbs and your internal link clicks go up—because readers stay to find the answer they didn’t know how to search for.

Reply with the last keyword cluster that didn’t convert the way you expected—we’ll feature operator audits in an upcoming piece.

2 August 2026
WordPress multisite user roles: when admin doesn’t mean super admin
Photo: Matinbeigi via Wikimedia Commons (CC BY-SA 4.0)

WordPress multisite introduces a permissions layer most solo operators don’t expect: the distinction between a site administrator and a network super admin. If you’ve ever granted someone “admin” access and watched them fail to install plugins, or wondered why a client can’t manage billing when they “own” their site, this is why.

The role split isn’t a bug. It’s architectural. But it catches people off guard because single-site WordPress collapses these into one role, and most multisite documentation assumes you’re running a university network, not a cluster of client sites or a portfolio of properties you manage alone.

What each role actually controls

A site administrator has full control over content, users, themes, and settings within a single site on the network. They can publish posts, moderate comments, manage menus, assign roles to other users on that site, and switch between themes—if you’ve allowed theme switching at the network level.

They cannot install or activate plugins, install new themes, edit PHP files via the theme editor (even if it’s enabled), access the network admin dashboard, or create new sites. Those permissions belong exclusively to the super admin.

A super admin controls the entire network. They can create and delete sites, install and network-activate plugins, upload themes, manage network-wide settings, and promote or demote users across all sites. They also inherit site-level admin privileges on every site in the network, whether explicitly assigned or not.

If you’re running a multisite as a solo operator—maybe you manage five niche content sites under one installation, or you host client projects—you’re almost certainly the only super admin. Everyone else, including clients who “own” their site, is a site administrator at best.

When role confusion breaks workflows

The most common issue: a client or collaborator reports they can’t install a plugin you told them to use. You check their account, see “Administrator” next to their name, and assume the platform is broken. It’s not. They’re a site admin, not a super admin, and multisite doesn’t let site admins touch plugins by default.

The second surprise: user management. A site administrator can add users and assign roles on their site, but they can’t remove a user from the network entirely. If someone needs to be deleted—not just demoted or removed from one site—that’s a super admin task. This creates cleanup debt if you don’t audit regularly.

The third trap: billing and domain mapping. If you’re using a plugin like Mercator or a host-managed domain mapping tool, site admins often can’t change the primary domain, even if they “own” the site contractually. That permission usually requires super admin access or a custom capability you’ve explicitly granted.

How to audit and delegate access correctly

Start by listing every user with super admin privileges. In the network admin dashboard, go to Users → Super Admins. If you see names you don’t recognise, or former contractors still listed, remove them immediately. Super admin is an all-or-nothing role; there’s no way to grant partial network control without a custom plugin.

For site-level access, go to each site’s dashboard and review Users → All Users. Confirm that client administrators are scoped to their site only. If someone needs plugin installation rights but shouldn’t control the whole network, you have three options:
- Pre-install and network-activate the plugins they need, then let them configure settings as a site admin
- Use a plugin like Multisite Plugin Manager or User Role Editor to grant granular capabilities—like install_plugins—to site admins on a per-site basis
- Promote them to super admin temporarily, have them complete the task, then demote them (risky, not recommended for clients)
Most operators choose option one. It’s cleaner and avoids the support overhead of explaining why a client can activate a plugin but not delete it from the network.

One non-obvious behaviour: user registration on multisite

If you enable open registration at the network level (Network Admin → Settings → Allow new registrations), users who sign up are added to the network, not automatically to any specific site. They exist in the user table but have no role and no dashboard access until a site admin or super admin assigns them to a site.

This confuses operators who expect self-service membership sites. A user registers, receives a confirmation email, logs in, and sees a blank dashboard with no menu items. They’re authenticated but not authorised anywhere. You have to assign them to a site and give them a role—Subscriber, Contributor, whatever fits—before they see content or functionality.

If you’re running a paid community or course platform on multisite, automate this with a membership plugin that handles role assignment on purchase. Don’t rely on WordPress’s default registration flow.

Multisite user roles aren’t complicated once you internalise the two-tier structure. But if you’re migrating from single-site WordPress or onboarding a client who expects full control, set expectations early. “Administrator” means something different here, and assuming otherwise costs you support time you don’t have.

One Two Three Send publishes operator-focused breakdowns like this one every day. Subscribe to catch the next deep dive on the tools and infrastructure that actually matter.
1 August 2026
Google Analytics 4 custom event parameters: the 25-limit nobody explains
Photo: Ajiro Shinpei via Wikimedia Commons (CC BY-SA 4.0)

Google Analytics 4 gives you almost unlimited flexibility to track custom events. You can fire anything: lead_form_submit, coupon_applied, video_watched. But there’s a hard constraint most solo operators don’t discover until it’s too late: GA4 only indexes 25 custom event parameters per property.

After that, new parameters still get logged in the raw event stream—but they won’t appear in standard reports, Explore, or Looker Studio. You can’t dimension or filter by them. They’re effectively invisible unless you’re pulling BigQuery exports, which most small operators aren’t.

This isn’t a bug. It’s a design decision Google made to keep the product performant. But it catches people by surprise because GA4’s interface doesn’t warn you when you’re approaching the limit, and old Universal Analytics didn’t have this restriction in the same way.

How the 25-parameter limit actually works

When you send a custom event to GA4—say, newsletter_signup with parameters like source, landing_page, referrer, and email_domain—those parameters need to be manually registered as custom dimensions in the GA4 admin panel before they show up in reports.

GA4 gives you:
- 25 custom dimensions (event-scoped)
- 25 custom dimensions (user-scoped)
- 50 custom metrics (numeric values)
Event-scoped dimensions are what most operators burn through first. These are things like button_label, video_title, product_category—anything that describes a single interaction.

Once you hit 25 event-scoped dimensions, you’re done. You can’t add more without archiving an existing one. And archiving doesn’t free up the slot—it just stops collection. Historical data stays, but the dimension becomes read-only.

What breaks when you hit the ceiling

Let’s say you’re tracking newsletter signups across six different lead magnets. You’ve been sending lead_magnet_name as a parameter for months. Then you launch a new sponsored post tracking setup and add five more parameters: sponsor_name, placement_type, cta_variant, reader_segment, and content_topic.

You go to GA4 Explore to build a report. The new parameters don’t show up in the dimension picker. You check the raw event in DebugView—it’s firing correctly. The data is being sent. But it’s not indexed, so it’s not queryable.

Here’s what you lose:
- You can’t segment audiences by that parameter
- You can’t build Explore reports around it
- You can’t use it in Looker Studio dashboards
- You can’t create conversion funnels that filter by it
The only workaround is BigQuery, which requires a GA4 360 subscription (starting at $50,000/year) or a manual export setup most indie operators won’t bother with.

How to plan your parameter budget

The fix isn’t technical—it’s editorial. You need to treat custom dimensions like a finite resource and plan what you track before you start sending events.

Start by auditing what you’re already using. Go to Admin > Data display > Custom definitions in GA4. You’ll see a list of every registered dimension and metric. Count them. If you’re above 20, you’re in the danger zone.

Then ask: Which of these dimensions do I actually query? Most operators register parameters “just in case” and never look at them again. Archive anything you haven’t used in a report in the last 90 days.

For new tracking, consolidate where you can. Instead of separate parameters for lead_magnet_name, lead_magnet_category, and lead_magnet_format, use a single lead_magnet_id and map it to a lookup table in your reporting layer. Instead of tracking button_color, button_size, and button_position separately, combine them into one button_variant string like blue_large_sidebar.

This isn’t elegant, but it works. And it keeps you under the limit.

The non-obvious tip: namespace your parameters early

If you’re starting fresh or still have slots available, prefix your custom parameters by category. Use form_name, form_step, form_source instead of generic names like name, step, source. It makes your dimension list easier to scan, reduces the chance of accidental overwrites, and helps you spot redundant tracking before you register a new dimension.

And when you do hit the limit? Don’t panic and start archiving things randomly. Export your current Explore reports first, note which dimensions they depend on, and only archive parameters that aren’t load-bearing.

Want more breakdowns like this? Reply with the analytics edge case that’s been tripping you up—we’ll cover it in a future issue.
1 August 2026
Affiliate link management: when spreadsheets cost you money

Photo by Gorilla ROI Data Connector on Unsplash

If you run affiliate links across your content—newsletter, blog, social—you probably started with a spreadsheet. Product name, affiliate URL, commission rate, maybe a notes column. It works until it doesn’t.

The breaking point isn’t volume. It’s versioning. Affiliate programs change their URLs, update terms, expire cookies faster, or shut down entirely. Your spreadsheet doesn’t tell you when a link dies. You find out three months later when a reader replies asking why your checkout link 404s.

What breaks first

Spreadsheets fail at three things: link rot detection, historical performance, and propagation speed.

Link rot is silent. An affiliate program migrates to a new domain, updates their tracking parameter structure, or sunsets a product SKU. Your old link still resolves—it just doesn’t credit you. Unless you’re manually testing every link monthly, you won’t know.

Historical performance matters when you’re deciding what to promote next quarter. A spreadsheet can log clicks if you’re using a link shortener with analytics, but matching clicks to actual conversions requires stitching together your shortener dashboard, the affiliate network backend, and your own notes. Most operators give up and optimize for vanity metrics instead.

Propagation speed is the operational bottleneck. You update a link in your spreadsheet, then you have to manually find and replace it across every past article, email archive page, and pinned social post. If the link appears in 40 places, you’re burning an hour. If you skip the older posts, you’re leaving dead links live.

When a link manager pays for itself

Dedicated affiliate link management tools—Pretty Links, ThirstyAffiliates, Lasso—charge $10 to $30/month. The ROI threshold is straightforward: if one broken link costs you more than one month’s subscription in lost commissions, the tool pays for itself.

For a solo operator earning $500/month in affiliate revenue, a single high-value conversion is worth $50 to $200 depending on the program. One missed sale covers six months of tooling.

The feature that matters most isn’t the link cloaking or the pretty dashboard. It’s automatic redirect updating. You edit the destination URL once in the tool’s backend; every instance of that short link across your entire site updates instantly. No find-and-replace. No archaeology through old posts.

Link health monitoring is the second-order benefit. Tools like Lasso ping your affiliate URLs weekly and flag 404s or redirects that don’t resolve to the expected domain. You get an alert, fix it, move on. The alternative is discovering the problem when a reader emails you or when you notice commission drops in your next payout statement.

The spreadsheet-plus-shortener hybrid

If you’re not ready to pay monthly, the middle path is a spreadsheet plus a custom domain short link service. Rebrandly’s free tier gives you 500 branded links and click tracking. You store the short link in your spreadsheet, paste that short link everywhere, and update the destination URL in Rebrandly when the affiliate program changes.

This works if you have fewer than 50 active affiliate relationships and you’re disciplined about logging every new link. It breaks when you forget to add a link to the sheet, or when you need to bulk-edit links by category (“update all Amazon links to the new Associate ID”).

The hidden cost is context switching. Every time you create a new affiliate link, you’re opening three tabs: the affiliate dashboard to generate the URL, Rebrandly to shorten it, and your spreadsheet to log it. That’s 90 seconds per link. If you’re adding 10 links a week, that’s 15 minutes weekly—13 hours a year—on administrative overhead.

What to do Monday

Audit your last 90 days of affiliate links. Open your spreadsheet, click every URL, and verify it resolves to the correct product page with your tracking parameter intact. If more than 10% are broken or redirect incorrectly, you have a link rot problem worth solving.

If you’re earning less than $200/month in affiliate revenue, stay with the spreadsheet but set a calendar reminder to re-check links quarterly. If you’re above $500/month or managing more than 30 active programs, trial a link manager for one month and measure time saved on link updates.

The goal isn’t perfect tracking. It’s reducing the lag between when a link breaks and when you notice. Every day a broken link stays live is a day you’re sending traffic you can’t monetize.

One Two Three Send covers the tools and tactics solo operators actually use to run content businesses. Subscribe to get one focused article daily—no filler, no fluff.

1 August 2026
Canva’s Magic Resize: when aspect-ratio automation saves time vs. breaks it
Photo: Bystronic Corporate Communications via Wikimedia Commons (CC BY-SA 4.0)

Canva’s Magic Resize feature promises to turn one social graphic into ten platform-ready versions with a single click. You design once at 1080×1080, hit resize, select Instagram Story + LinkedIn post + Pinterest pin, and the tool redistributes text boxes, images, and spacing to fit each canvas.
It works exactly as advertised—when your design is simple. Two text blocks, a centered logo, and a background gradient? Magic Resize handles it cleanly. But stack five text layers with custom positioning, overlay transparent PNGs, or use grouped elements with manual kerning, and you’ll spend more time fixing the output than you would rebuilding from scratch.
How Magic Resize decides what moves
Magic Resize uses a hierarchy system. Canva prioritizes foreground elements (text boxes, stickers, uploaded images) over background layers. When you switch from a 1:1 square to a 9:16 Story canvas, the tool expands the vertical space and redistributes elements top-to-bottom based on their original layering order.
Text boxes resize proportionally by default—font size stays fixed, but line breaks reflow to fit the new width. Images scale to maintain aspect ratio. Background fills stretch to cover the entire canvas without cropping.
The problem: Canva doesn’t understand visual hierarchy the way you do. If your Instagram square placed a call-to-action at the bottom because that’s where the eye naturally lands in that format, Magic Resize will keep it at the bottom of a tall Story canvas—where it’s offscreen until someone swipes up.
Grouped elements break the logic entirely. If you’ve grouped a text box with a shape to create a custom badge, Magic Resize treats the group as a single object and scales it proportionally. A badge that was 15% of your square design becomes 15% of your Story height—often way too large or positioned awkwardly.
When to use it vs. when to duplicate manually
Magic Resize works best for templated content where layout consistency matters more than pixel-perfect positioning. Weekly quote graphics, announcement cards, course promotional images—anything with one dominant message and minimal layering.
Skip it when:
- Your design uses more than four text layers with custom positioning
- You’ve applied manual spacing or alignment that depends on the original canvas shape
- Grouped elements make up more than 30% of the design
- You’re resizing from a horizontal format (16:9 YouTube thumbnail) to vertical (Pinterest pin)—the text reflow often breaks readability
For complex designs, duplicate the original file instead. Canva lets you copy an entire design and manually adjust the canvas dimensions. You’ll spend three minutes repositioning elements yourself, but you’ll have full control over what moves and what scales.
The non-obvious tip: pre-size your text boxes
Most operators let Canva auto-size text boxes—you type, the box expands to fit, and you drag it into place. That approach breaks Magic Resize because the tool doesn’t know which dimension (width or height) you prioritized.
Instead, manually set text box dimensions before you type. Click the text box, drag it to a fixed width (say, 800 pixels on a 1080-wide canvas), then type your headline. Lock the width in place. When Magic Resize redistributes the layout, it preserves your width ratio and reflows text vertically—keeping line breaks predictable.
Same goes for images. If you upload a logo or product shot, resize and position it first, then lock the layer. Magic Resize respects locked elements and works around them, giving you anchor points that stay consistent across formats.
Pricing and access
Magic Resize is available on Canva Pro ($15/month for solo users, $30/month for up to five team members) and Canva for Teams plans. The free tier doesn’t include it—you’ll need to manually duplicate and resize instead.
If you’re publishing to three or more platforms weekly and your designs stay simple (under five layers, minimal grouping), the feature pays for itself in saved time. If your graphics are layout-heavy or require tight control over spacing and alignment, the Pro subscription still has value for the template library and Brand Kit features, but Magic Resize won’t be the reason you renew.
Want more tool breakdowns like this? Subscribe to One Two Three Send and get operator-focused features, comparisons, and how-tos every week—no fluff, just the details that matter when you’re running a content business solo.
31 July 2026
WordPress transactional email: when wp_mail() fails silently
Photo by Mariia Shalabaieva on Unsplash

WordPress sends password resets, comment notifications, and form submissions through a single PHP function: wp_mail(). It works—until it doesn’t. And when it breaks, you usually won’t know.

There’s no delivery confirmation, no bounce handling, no retry logic. The function returns true if it handed the message to your server’s mail transport. What happens after that is invisible.

For solo operators running contact forms, membership sites, or order confirmations, silent email failure costs conversions and creates support overhead. Here’s when the built-in system breaks, what to replace it with, and how to make the switch without touching every plugin individually.

When wp_mail() actually breaks

WordPress uses your server’s local mail transfer agent by default—usually Sendmail or Postfix. Shared hosting providers throttle or block outbound SMTP to prevent spam. Budget VPS instances often ship with no MTA configured at all.

Even when mail leaves your server, deliverability suffers. Your domain lacks proper SPF and DKIM records for server-originated mail. Gmail and Outlook route it to spam or reject it outright. You’ll never see the bounce.

Common failure scenarios:
- Password reset emails never arrive, users assume the form is broken
- WooCommerce order confirmations vanish, customers contact support
- Gravity Forms or Contact Form 7 submissions disappear after the success message
- Comment reply notifications stop working, engagement drops
The function still returns true. WordPress has no idea delivery failed.

Replace wp_mail() with an SMTP plugin or API bridge

You need to route WordPress mail through a transactional email service with proper authentication and delivery tracking. Two approaches work:

SMTP plugins reconfigure wp_mail() to connect to an external mail server. WP Mail SMTP (free) and Post SMTP (free) both support Gmail, SendGrid, Mailgun, and Amazon SES. You add credentials, test a message, and every plugin that calls wp_mail() automatically routes through the new transport.

Setup takes ten minutes. The catch: SMTP connections can time out under load, and you’re still managing API keys in the WordPress admin.

API-based plugins replace the SMTP handshake with HTTP calls. Postmark’s official plugin (free) and Mailgun’s plugin both inject an API bridge. Delivery is faster, retries are automatic, and you get real bounce logs in the service dashboard.

For solo operators sending under 1,000 transactional emails per month, Postmark’s free tier covers password resets and form notifications. Once you cross 1,000 sends, pricing starts at $10/month for 10,000 emails—still cheaper than the support time spent debugging silent failures.

One non-obvious configuration detail

After you install an SMTP or API plugin, set a dedicated “From” address and verify it with your email service. Most plugins default to [email protected], which triggers spam filters if that address doesn’t exist or lacks DNS records.

Create [email protected] or [email protected] as a real mailbox (or alias), add it to your transactional service’s verified sender list, and configure the plugin to use it. This single step fixes 80% of residual deliverability problems after switching away from wp_mail().

Also: enable logging in the plugin settings. WP Mail SMTP and Post SMTP both include email logs that show exactly what fired, when, and to whom. When a customer says they didn’t get a receipt, you’ll know whether it was sent.

When to stay with wp_mail()

If you’re running a single-author blog with comments disabled and no forms, the built-in function is fine. You’re only sending password resets to yourself, and you’ll notice if those break.

But the moment you add a contact form, membership plugin, e-commerce checkout, or any user-triggered email, the risk shifts. Silent failure becomes a business problem, not a technical curiosity.

Switching to a transactional service doesn’t require a developer. Install a plugin, add an API key, send a test. If the test arrives and your logs confirm it, you’re done.

What’s breaking on your WordPress site right now? Hit reply and tell us which email mystery you’ve been ignoring. We’ll cover it in a future piece.
31 July 2026
AI writing assistants charge per seat—when to share logins instead

Photo by Emil Karlsson on Unsplash

AI writing tools have adopted SaaS pricing: one seat, one user, one monthly fee. Add a second person—a contract editor, a VA who schedules posts, a designer who needs context—and you’re suddenly paying double.
For solo operators who occasionally collaborate, that math doesn’t work. You’re not running a newsroom. You don’t need role-based permissions, audit logs, or centralised billing. You need someone to proofread a draft on Tuesday and disappear until next month.
Most AI platforms don’t want you sharing logins, but the pricing gap between individual and team plans creates a structural mismatch for small operations. Here’s when sharing credentials makes operational sense, and when you actually need to pay for seats.
When shared logins work fine
If your collaborator needs access fewer than five times a month, and you’re not working in the tool simultaneously, a shared login handles it. One operator I talked to shares her Claude account with a contract editor who fact-checks newsletter drafts twice a week. The editor logs in, opens the shared thread, leaves comments, logs out. Total time: under an hour per session.
The workflow constraint is synchronous conflict. If you’re both drafting in the same thread at 9am, someone’s work gets overwritten. For async handoffs—”I drafted this, you edit it by Thursday”—shared access works.
Most AI tools don’t enforce device limits or concurrent session blocks on individual plans. ChatGPT Plus, Claude Pro, and Jasper’s Starter tier all allow login from multiple devices without flagging the account. They assume you’re switching between your laptop and phone, but the technical guardrail is the same.
When you need to pay for seats
Team plans make sense when you need simultaneous access, conversation history separation, or compliance coverage. If you’re working in the tool at the same time—co-editing a product launch sequence in real time—shared credentials break down. You’ll overwrite each other’s changes or lock each other out mid-session.
Conversation history is the other friction point. Shared logins mean shared threads. If your VA is using the account to draft social captions while you’re debugging a landing-page headline, your chat history turns into an unnavigable mess. Finding yesterday’s draft means scrolling past twenty unrelated threads.
Some platforms offer workspace separation on team plans. Jasper’s Business tier gives each user isolated project folders. ChatGPT Team lets you create shared spaces without bleeding personal threads into the company view. If you’re collaborating weekly or more, that structure is worth the seat cost.
Compliance matters if you’re handling customer data or operating in a regulated niche. Shared logins muddy accountability—your Terms of Service violation could be your editor’s mistake, but the platform sees one account. Team plans with per-user audit trails give you a paper trail if something breaks.
Pricing breakpoints that matter
Claude Pro costs $20/month for individuals. Claude Team starts at $30/month per seat, minimum two seats, so $60/month total. If your editor logs in twice a month, you’re paying $40/month for convenience features you don’t use.
ChatGPT Plus is $20/month individual, $25/user/month for teams (minimum two users, so $50/month). The team plan adds shared conversation history and admin controls. For occasional collaboration, that’s $30/month you’re spending to avoid saying “log out when you’re done.”
Jasper starts at $39/month for individuals (called Creator), then jumps to $99/month for Teams. The gap is wide enough that most solo operators share logins until they’re collaborating daily.
The tipping point isn’t about budget—it’s about friction cost. If you’re losing twenty minutes a week to login coordination, thread archaeology, or overwritten drafts, the team plan pays for itself. If you’re handing off async once a week with zero conflicts, keep sharing.
The tool doesn’t care—until it does
AI platforms bury account-sharing language in their Terms of Service, but enforcement is rare for individual-tier plans. They’re optimised to catch resellers and abuse (one account serving a dozen users), not a solo operator splitting access with a part-time editor.
That said, if you’re logging in from different cities on the same day, or running hundreds of queries in parallel, you risk a flag. The platform sees usage patterns, not intent. Keep activity reasonable—if it looks like one person’s workload, you’re fine.
One non-obvious risk: password resets. If your collaborator changes the password and forgets to tell you, you’re locked out mid-project. Set up a shared password manager (1Password, Bitwist) so credentials live in one auditable place, not a text thread.
If you’re already running a team plan for other tools, check whether your AI assistant offers bundled pricing. Some operators I know share a ChatGPT Team account purely because they were already paying for shared Notion and Figma—it’s one less login to manage.
Want more operator-level breakdowns of tools, pricing, and workflows that actually matter? Subscribe to One Two Three Send for weekly deep-dives into the small decisions that compound.

Heads up — some links in this article are affiliate links. If you sign up through them, we may earn a small commission at no extra cost to you. We only recommend tools we use ourselves.

31 July 2026
Newsletter archive search: when readers can’t find what you wrote

Photo: Unknown via Wikimedia Commons (Public domain)

A reader emails: “I know you wrote about WordPress caching last month—where is it?” You know you sent it. You can see it in your sent folder. But when they search your archive page, nothing turns up.
Archive search is one of those features newsletter platforms advertise but rarely explain. Some index subject lines only. Others index the first 200 characters. A few index the full body—but only if you’re on a paid tier. And almost none tell you which approach they use until you test it yourself.
Here’s what actually gets indexed on the platforms solo operators use most, and what to do when search fails.
What gets indexed on each platform
Beehiiv indexes the full email body on all tiers, including free. Search works across subject lines, body text, and author names if you’ve enabled bylines. The index updates within a few minutes of sending. One caveat: if you use custom HTML blocks, only plain text inside those blocks gets indexed—images, buttons, and styled div containers are ignored.
Substack indexes subject lines and body text, but search results prioritize exact matches in titles first. If your subject line was vague (“Issue #47”) and the meat of the topic is buried in paragraph three, it won’t surface until a reader scrolls past a dozen other posts. Substack’s archive search also doesn’t support Boolean operators—no “AND,” “OR,” or quoted phrases.
ConvertKit offers archive search only if you’ve enabled the public archive feature, which is off by default. When enabled, it indexes subject lines and the first 500 characters of body text. Anything below the fold in your email—your detailed how-to, your step-by-step breakdown—doesn’t get indexed. If you bury the lede, readers won’t find it.
MailerLite doesn’t offer built-in archive search at all. The public archive page displays a reverse-chronological list with subject lines and send dates. Readers have to use browser find (Cmd+F) and hope the subject line matches what they remember. For a catalog of 50+ emails, that’s borderline unusable.
When to use tags and categories instead
If your archive grows past 30 posts, search alone won’t save readers. They need filters—by topic, by format, by content type. Beehiiv lets you tag posts and surface those tags as filters on your archive page. Substack offers Sections, which let you split your newsletter into topic-based streams (each with its own archive and RSS feed).
ConvertKit doesn’t support tags or categories in the public archive view. If you need taxonomy, you’ll need to maintain a separate landing page with manual links—or migrate to a platform that supports it natively.
The non-obvious move: if you publish consistently on a topic (say, WordPress hosting or AI prompts), create a dedicated landing page that lists all related posts with short descriptions. Link to it from your welcome email and your site nav. That page becomes your real archive. The platform’s built-in search becomes a backup.
What breaks when readers search from mobile
Most newsletter platforms serve the same archive page to mobile and desktop, but mobile browsers handle search differently. Safari on iOS doesn’t support in-page search widgets that rely on JavaScript—if your platform uses a custom search bar (instead of a plain HTML form), iOS readers get a broken experience. They tap the search icon, nothing happens, and they leave.
Beehiiv and Substack use standard HTML forms, so mobile search works. ConvertKit’s archive search widget relies on JavaScript and fails silently on older mobile browsers. MailerLite, again, offers no search at all.
Test your archive page on an actual phone, not just a resized browser window. Open it in Safari, Chrome, and Firefox mobile. Try searching for a post you know exists. If the search bar doesn’t respond or returns zero results for a term you know is there, your readers are hitting the same wall.
When to embed a third-party search tool
If your archive has 100+ posts and your platform’s search isn’t cutting it, you can embed a third-party site search tool. Algolia offers a free tier for up to 10,000 searches per month. You’ll need to generate a JSON feed of your archive (most platforms support RSS; convert it to JSON with a script or a tool like Feed43), push it to Algolia’s index, and embed their search widget on a custom landing page.
This works if you host a separate website alongside your newsletter (common for operators who publish on Substack but maintain a WordPress site for SEO). It doesn’t work if you rely solely on the platform’s hosted archive—Substack and Beehiiv don’t let you inject third-party JavaScript into their archive pages.
The simpler fix: write better subject lines. If every email is titled “Weekly Update” or “Issue #23,” no search tool will help. Use the subject line to signal the topic clearly. “WordPress object caching: Redis vs. memcached” beats “Performance tips” every time.
Want more tool breakdowns like this? Subscribe to One Two Three Send and get one operator-focused deep-dive every day—no fluff, just what actually works and what breaks.

Heads up — some links in this article are affiliate links. If you sign up through them, we may earn a small commission at no extra cost to you. We only recommend tools we use ourselves.

30 July 2026
Notion databases as CRMs: when it breaks and what to use instead

Photo: Saumya Singh 06 via Wikimedia Commons (CC BY-SA 4.0)

Notion databases feel like the perfect CRM when you’re managing your first dozen sponsor contacts or affiliate relationships. Drag a card, add a relation, filter by status—it’s visual, it’s flexible, and it’s already part of your workspace.
Then you hit 200 contacts. Or you need to send 40 follow-up emails in one afternoon. Or you want to see which sponsors opened your pitch. That’s when the cracks show.
Notion wasn’t built to be a CRM. It was built to be a database interface inside a document editor. The distinction matters more than most solo operators realise until they’re deep enough in that migrating feels painful.
Where the Notion database model fails
The first breaking point is bulk actions. Notion lets you select multiple database entries and change a single property—status, tag, date. But if you need to update five fields across 30 records, or send templated emails to a filtered segment, you’re clicking into each row one by one.
There’s no native way to trigger an action when a property changes. You can’t auto-send a follow-up email when a sponsor moves to “Negotiating,” or log a timestamp when a contact replies. Zapier can bridge some of this, but you’re bolting automation onto a tool that doesn’t expose the hooks a real CRM would.
Search breaks down fast. Notion’s database search works within a single view at a time, and it’s not full-text across linked databases. If you’ve split your sponsor contacts, media kit versions, and pitch history into separate tables—like Notion best practices suggest—you can’t search across all three and get a unified contact timeline.
Permissions get messy when you bring on a contractor or co-founder. Notion’s sharing model is page-level. You can’t give someone access to sponsor contacts without also giving them access to the parent page, which might include financials, drafts, or unrelated projects. Real CRMs let you control visibility at the record or field level.
When Notion is still the right call
If you’re managing fewer than 100 contacts, aren’t doing daily outreach, and don’t need email tracking or automation, Notion’s database is fine. It’s especially good if your CRM needs overlap heavily with project documentation—keeping pitch decks, negotiation notes, and contract PDFs in the same workspace where you track the relationship.
Notion also works when your “CRM” is really just a lead list. If you’re sourcing potential sponsors from a directory, tagging them by fit, and handing off the actual outreach to a platform like Mailshake or Lemlist, Notion can serve as the pre-CRM layer without strain.
What to use when you outgrow it
Most solo operators don’t need Salesforce. You need something that handles contact records, email sequences, and basic pipeline tracking without a multi-week setup.
Brevo (formerly Sendinblue) includes a lightweight CRM alongside its email platform. You get contact records, deal pipelines, and the ability to trigger email automation when a deal stage changes. Pricing starts free for up to 300 emails per day, and the CRM features are included even on the free tier. It’s a good fit if your CRM is primarily for sponsor or affiliate outreach and you want email tracking built in.
Airtable sits between Notion and a real CRM. It’s still a database tool, but it has better bulk-edit controls, more powerful filtering and grouping, and a cleaner API if you’re connecting it to Zapier or Make. The interface feels more like a spreadsheet than a document, which some operators prefer once contact volume climbs. The free tier caps at 1,000 records per base.
Streak lives inside Gmail and turns your inbox into a CRM. Each contact thread becomes a pipeline card. You can see open rates, set reminders, and log notes without leaving your email client. It’s $15/month for the solo plan and works well if most of your relationship management happens over email. The tradeoff: no standalone contact view outside Gmail, so reporting and bulk edits are limited.
Migration friction is real—plan for it early
The hardest part of moving off Notion isn’t picking the next tool. It’s extracting your data without breaking relationships between records. Notion’s CSV export flattens relations into plain text, so if you’ve linked sponsor contacts to pitch history or media kits, those connections don’t survive the export.
If you’re at 80 contacts and starting to feel the friction, that’s the time to move—not at 300 when you’ve got two years of notes embedded in linked databases. Set aside an afternoon, export your main contact table, and rebuild the essentials in the new tool. Don’t try to migrate every field or historical note. Carry forward active contacts, current pipeline deals, and the last six months of interaction history. Archive the rest in Notion as read-only reference.
The goal isn’t to find the perfect CRM. It’s to use a tool that doesn’t force you to work around its limitations every time you need to do something twice.
Reply with the tool you’re using to manage sponsors, affiliates, or client contacts—I’m tracking what solo operators are actually reaching for in 2026.

Heads up — some links in this article are affiliate links. If you sign up through them, we may earn a small commission at no extra cost to you. We only recommend tools we use ourselves.

30 July 2026

► Necessary Cookies Always Active

Necessary cookies enable essential site features like secure log-ins and consent preference adjustments. They do not store personal data.

► Functional Cookies Remark

Functional cookies support features like content sharing on social media, collecting feedback, and enabling third-party tools.

► Analytical Cookies Remark

Analytical cookies track visitor interactions, providing insights on metrics like visitor count, bounce rate, and traffic sources.

► Advertisement Cookies Remark

Advertisement cookies deliver personalized ads based on your previous visits and analyze the effectiveness of ad campaigns.