Three PR strategies from 2014 still hold in 2026: keep content simple, build for mobile, and win back shoppers who leave without buying. StatCounter reported that mobile and tablet devices passed desktop worldwide in October 2016, 51.3% to 48.7%. The Baymard Institute puts average cart abandonment at 70.22%. Each strategy now has an AI-search version, and a brand that skips it misses the shortlist an assistant gives a shopper.
Originally published January 8, 2014. Updated October 2026.
What did the January 2014 post predict?
The January 2014 post named three strategies that would shape PR for the rest of the decade. The first was to keep content simple and clean, because readers were overwhelmed by information and wanted concise, useful pages fast. The second was to build for mobile. The third was to retarget the visitors who leave without buying, since a store's competitors can reach them too.
Twelve years of evidence support all three calls, with one correction to the original timeline. An earlier version of this page said mobile traffic passed desktop in 2014. StatCounter's data shows that crossover happened worldwide in October 2016, so the call was early, but the direction was right.
| 2014 strategy | What the evidence shows | 2026 version |
|---|---|---|
| Keep content simple and clean | Short, structured pages are easier for readers and software to use | Pages written so an AI assistant can quote one passage |
| Build for mobile | Mobile and tablet passed desktop worldwide in October 2016 | Content built for AI answer engines as the next device layer |
| Retarget non-buyers | About 70% of carts are abandoned, per Baymard | Consistent brand facts wherever shoppers compare options |
Did "keep it simple" hold up?
Keeping content simple held up because clear pages serve two audiences at once: busy readers and the software that summarizes the web for them. A page with one claim per section, named entities and exact numbers is faster to read, and an AI assistant can lift a single passage from it without distortion.
Why it works: an AI assistant builds an answer from passages it can find and attribute to a named source. A page that states its point in the first sentence under each heading gives the assistant a passage that stands alone. In 5W's May 2026 Venture Capital AI Visibility Index, a16z.com appeared in 7.1% of retrieved responses, the only venture-firm domain in the top 10 sources. Clear, regularly published pages gave assistants more to quote.
The 2014 advice was about readers' time. The 2026 version adds a second test: read each paragraph with everything above it hidden, and check that it still makes sense. Replace pronouns with the names of the people, products and places involved. That habit costs nothing and helps both audiences.
When did mobile overtake desktop, and what did it change?
Mobile overtook desktop worldwide in October 2016, when StatCounter measured 51.3% of internet usage on mobile and tablet devices against 48.7% on desktop. Desktop still led in the United States at 58%, and StatCounter's chief executive told small businesses to check that their sites worked on phones.
Search engines followed the audience. According to StatusCake's summary, Google announced in November 2016 that it was running experiments to make its index mobile-first, three days after the StatCounter report. A site that worked only on desktop risked a lower rank, so mobile-ready pages became a ranking requirement and not only a courtesy.
The same pattern is repeating with AI. Capgemini's 2026 consumer report, based on 12,000 consumers in 12 countries, says AI-driven tools are reshaping what shoppers consider acceptable, valuable and worth paying for. The brands that rebuilt for phones early kept their search visibility. The brands that restructure their content for AI assistants now will have the same head start.
Why retarget shoppers who leave without buying?
Retargeting works because most shoppers leave without buying, and many of them are still deciding. The Baymard Institute, which combined 50 studies, reports an average cart abandonment rate of 70.22%. Baymard also notes that a large share of abandonment is natural, since people window shop, compare prices and save items for later.
That is the audience the 2014 post described. A shopper who leaves is not a lost customer, and the brand that reaches that shopper again with a useful message often wins the sale. Baymard estimates that better checkout design alone could raise the conversion rate of an average large site by 35.26%, so some of that audience is recoverable on the brand's own pages.
The 2026 version moves the comparison step. Many shoppers now ask an AI assistant to compare options before they return to any store. If a brand's site, retailer listings and press coverage give different prices or claims, the assistant has no reliable answer to quote. Consistent facts across those sources play the role that repeat display ads once played: they keep the brand in front of the shopper at the moment of comparison. The post on citation share as a new market-share metric explains how to measure that presence.
How can a brand apply the three strategies this quarter?
A brand can apply the three strategies in one quarter by auditing its pages for clarity, testing them on phones and assistants, and checking its facts across sources. The table sets out the steps in order.
| Step | Action | Output |
|---|---|---|
| 1. Simplify | Rewrite each key page so every heading is followed by a self-contained answer | Pages an assistant can quote without context |
| 2. Test on devices | Load the pages on a phone and ask two AI assistants a question the page should answer | A list of gaps between what the page says and what the assistants return |
| 3. Match the facts | Compare prices, claims and descriptions across the site, retailer listings and press coverage | One approved fact sheet and a fix list |
| 4. Reach non-buyers | Follow up with visitors who left, using a message that answers their likely question | A short email or ad sequence tied to the page they viewed |
| 5. Measure | Re-run the assistant questions each month and record which brands appear | A dated log of presence over time |
Run steps one to three before spending on step four. A follow-up message that sends shoppers to a confusing page wastes the second chance.
How do the three strategies work together?
The three strategies work as one sequence: clear pages earn the first look, mobile and assistant-ready formats put those pages where people ask, and follow-up reaches people who have not decided. Skipping any link weakens the chain. A clear page that loads slowly on a phone loses the visitor, and a fast page full of vague claims loses the assistant's quote.
Think of the sequence as a funnel with three questions. Can a reader understand the page in ten seconds? Can the page be reached on the device and in the tool the customer is using? Does the brand follow up when the customer leaves? A team that answers yes to all three has the 2014 playbook in working order.
What does the 2026 check look like in practice?
The 2026 check can be done in one working session per page and uses three tests: a read-aloud test, a phone test and an assistant test. Illustrative scenario: a regional furniture retailer picks its sofa-care guide, its delivery policy and its return policy as the three pages customers ask about most.
For the read-aloud test, a team member reads each heading and the sentence under it, with nothing else visible, and marks any sentence that needs earlier context. For the phone test, the team loads each page on a mid-range phone over a mobile connection and notes anything that is slow or hard to tap. For the assistant test, the team asks two AI assistants a plain question, such as "what is the return policy at this retailer?", and compares the answers with the page.
The differences become the fix list. If an assistant quotes an old return window, the retailer updates the page, its listings and any press mention that carries the old figure. The same method works for any brand, and it uses the same discipline the 2014 post asked for.
Which numbers show whether the strategies are working?
Three numbers show progress: the share of key pages that pass the read-aloud test, the mobile load and tap problems still open, and the number of assistant answers that match the brand's own facts. Track each one monthly, and keep the log dated so a later reader can see the trend.
Add one business number alongside them, such as repeat visits or recovered carts, so the work connects to revenue. Baymard's figures give a benchmark for the recovery side: a 70.22% average abandonment rate means even a small improvement in follow-up or checkout clarity reaches many shoppers. Report the result in plain language to executives, with the date of each measurement.
What risks come with copying a 2014 playbook?
The main risk is treating a twelve-year-old playbook as a guarantee. Display retargeting costs money on every impression, and tracking options differ by platform today. The principle, which is to stay useful to people who have not yet decided, holds. The tools change, so test each channel with your own numbers.
A second risk is over-simplifying. Clear writing does not mean thin content. A page still needs the evidence that makes a claim believable: a named source, a date and a number. The 2014 advice asked for concise pages, and it never asked for empty ones.
A third risk is chasing every new channel. Mobile took years to become the default, and AI assistants will take their own time to settle. Pick the two or three questions your customers ask most often, and make sure your pages answer them well before expanding.
Where does this post sit in the archive?
This post belongs to the 2014 to 2016 archive, which collects the early posts that this site's later research builds on. The full chronological record is in 23 Years of Communications Thinking, and the anchor events through-line shows how those early ideas connect to the current research program.
Reading the 2014 post next to this one shows what stayed constant. The tools changed from display ads and mobile sites to structured pages and AI assistants. The discipline stayed the same: be clear, be present where people look, and follow up with people who have not decided.
Conclusion: Keep the 2014 discipline and update the tools
The three 2014 strategies hold because they describe how people behave, not which platform they use. This month, pick your five most visited pages and rewrite each one so every heading is followed by a complete answer. Test them on a phone and in two AI assistants, fix any facts that differ across sources, and then reach the shoppers who left. Record the results, and repeat the test next quarter.
Last updated: October 2026.