44 growth hacks you can start this afternoon
TL;DR
Difficulty: Beginner
Most growth comes from fixing obvious problems, not from secret tactics. This list holds 44 fixes you can start in an afternoon: a reachable button on phones, real photos, fewer form fields, honest urgency, better page titles and internal links, answers in the first sentence for AI search, a deliberate robots.txt, a clean and lawful email list measured on clicks rather than opens, negative keywords, consented retargeting, and a comparison of what a customer is worth against what one costs. Set up conversion tracking with consent first, then change one thing at a time and keep a log.
What to read first
The first version of this list promised 53 growth hacks with confident impact figures, and most of those figures had no source at all. I checked every claim, dropped the tactics that were thin, out of date or in conflict with consent law, and kept 44 fixes you can start this afternoon. Each comes with what it solves and how to do it, and the figures that remain have a source you can open. A worked example for an invented Bristol tea shop shows how to compare what a customer is worth with what one costs.
Most growth comes from fixing what is obviously broken
The first version of this article promised 53 growth hacks, each with a confident impact figure: "15 to 30% more mobile conversions", "each form field removed lifts completion by 10%". When I went back to check where those numbers came from, most of them came from nowhere. They were the kind of figures that circulate from one listicle to the next until everyone believes them. So I rewrote the list. Tactics that were thin, out of date or in conflict with consent law are gone or folded into others, and the figures that remain have a source you can open. That leaves 44.
What has not changed is the idea. Most websites and marketing accounts do not need a secret tactic. They need someone to spend an afternoon fixing the obvious: the button nobody can find on a phone, the form that asks for a fax number, the ad that shows for searches nobody wanted, the newsletter sent from "noreply". Each fix below is something you can start in an afternoon. Some take ten minutes, a few (original research, a definitive guide) are an afternoon of planning for a longer job, and I say so where that is the case.
One honest warning before the list. A tactic is a guess until you measure it. If you change five things at once and sales rise, you will not know which change did it, and you will not know whether the rise was the change or simply a sunny week. So the order matters: first make sure you can measure (tactic 36 is about conversion tracking with consent), then change one thing at a time on the pages that carry the most money, and keep a short log of what you changed and when.
Website: where small leaks add up
Most websites do not lose sales to one big flaw. They lose them to a dozen small frictions, each of which makes a few visitors give up. These are the ones I check first on any site.
1. Keep the main action within reach on a phone
On a phone, the call to action sits at the top of the page and disappears the moment someone scrolls. By the time they have read enough to be convinced, the button is three screens away. A slim bar fixed to the bottom of the screen, carrying one button with the main action ("Book a call", "Add to basket"), solves this. Keep it narrow, make sure it never covers form fields or the cookie banner, and leave room for the phone's own gesture bar. On the sites I build it is standard, and it is one of the few changes worth doing without a test, because the cost of having it is close to zero.
2. Replace stock photos with real ones
Eyetracking research from Nielsen Norman Group found that people scrutinise photos of products and of real people who actually work at a company, and ignore generic stock photos of models and big decorative images used to "jazz up" a page.1 Swap the hero image and the "about" section first. A well lit phone photo of your actual workshop, team or product does more than a polished picture of strangers shaking hands, because it carries information: this business exists and these are the people in it.
3. Put proof next to the headline, and only real proof
A visitor deciding in the first few seconds whether you are legitimate will not scroll to the testimonials at the bottom. Put one line of proof directly under the headline: a star rating with the number of reviews, the logos of two or three recognisable clients who agreed to be named, or one short quote with a full name. Make sure every piece is genuine. In the UK, writing or commissioning fake reviews, or hiding that a review was incentivised, has been an automatically unfair commercial practice since 6 April 2025 under the Digital Markets, Competition and Consumers Act 2024.2
4. Cut every form field you do not need before first contact
Every field is a small reason to leave. Baymard Institute's checkout research found that the average checkout in 2024 had 11.3 form fields, while most sites need only 8, and that 17% of users had abandoned a purchase because the checkout was too long or complicated.3 Lead forms suffer the same way. Go through each field and ask whether you genuinely need it before the first conversation. Name, email and one open question are usually enough. The "company size" dropdown that sales asked for can be a question on the first call.
5. Use an exit popup, but only one and only on desktop
A popup that appears when the mouse heads for the close button gives a leaving visitor one more reason to stay: a useful guide, a sample, a newsletter with a clear promise. Keep it to one sentence and one field, make the close button obvious, and never show it twice to the same person. Avoid it on mobile, where there is no mouse to track and where Google's guidance warns against interstitials and dialogs that cover the content people came for.4 If you ask for an email address, the consent rules in tactic 27 apply.
6. Make the site faster where it matters
Speed is not a vanity metric. A study by Deloitte for Google, published on web.dev, found that a 0.1 second improvement across four mobile speed metrics lifted the share of retail visitors moving from a product page to the basket by 9.1%, and that retail shoppers spent 9.2% more.5 The quick wins are almost always the same: images exported at the size they are shown and in a modern format such as WebP or AVIF, lazy loading for images below the first screen, and removing scripts nobody uses any more (old chat widgets, three analytics tools doing the same job). Run PageSpeed Insights on the pages that carry the most revenue and start with the first item it lists.
7. Say what you do in one plain sentence
If a visitor cannot tell within a few seconds what you sell, who it is for and why it is better, they leave. A useful test: your headline should complete the sentence "I help [a specific group] get [a specific result]" without a pun or a metaphor in sight. Read your current headline to someone who does not know your business and ask them to repeat back what you do. If they cannot, rewrite it today. Clarity beats cleverness every time.
8. Make the main button the most obvious thing on the page
The eye follows contrast, space and direction. If your main button is the same colour as five other elements, nobody notices it. Give the primary action one colour that is used for nothing else, leave white space around it, and let photos of people face towards it rather than off the edge of the page. Check the contrast while you are there: text on buttons needs a contrast ratio of at least 4.5 to 1 against its background to meet WCAG 2.1 AA, which is also what makes it readable in sunlight.6
9. Trim the navigation
A menu with twelve items asks visitors to make twelve decisions before they make one. Group secondary pages under a heading, move legal and career pages to the footer, and keep the top level to the handful of places most visitors actually go (your analytics will tell you which). Give the primary action its own button style in the header so it does not compete with "About".
10. Use urgency only when it is true
A real deadline (the sale ends on Sunday, the course has twelve places, the price rises on 1 January) gives people a reason to decide now. A fake one destroys trust and, in the UK, is now unlawful in itself: falsely stating that a product will only be available for a limited time, or on particular terms for a limited time, to rush a decision is one of the practices banned outright by Schedule 20 of the Digital Markets, Competition and Consumers Act 2024.2 Countdown timers that restart when the page reloads fall squarely into that. If the deadline is real, say what it is and why.
11. Answer the payment worry before it is asked
Baymard's survey of why shoppers abandon a checkout puts "I didn't trust the site with my credit card information" third, at 19% of those who were not just browsing, behind extra costs (40%) and slow delivery (20%).7 For a lesser known shop, the fix is quick to try: show the payment providers people recognise (PayPal, Apple Pay, Klarna, the card schemes) next to the payment form, state the returns policy in one sentence, show the full price including delivery before the last step, and make a phone number or chat visible.
SEO: the basics that still decide rankings
None of these is new. All of them are still missing on a surprising share of sites, and each one can be fixed on your most important pages in an afternoon.
12. Rewrite the titles of your ten most important pages
The page title is the blue link in the search results, and it is the first sentence a searcher reads. Google says there is no limit on how long a title can be, but that the link is cut off to fit the device, and that keyword stuffing and boilerplate titles should be avoided, because Google may rewrite them.8 Give each important page its own title, put the subject first, include the word people actually search for, and make it something a human would want to click. Fifty or sixty characters is a sensible working length, not a rule.
13. Write the description as if it were an advert
Google mostly builds the snippet under the title from the page itself, but it does use the meta description when that describes the page better than the page's own text.9 It is not a ranking factor, but it affects whether people click. Write one for each important page: what the page offers, for whom, and why this one, in around 150 characters. Avoid repeating the title and avoid one description copied across fifty pages.
14. Link your pages to each other, with words that describe the target
Google discovers and understands pages largely through links, and it recommends anchor text that says what the target page is about, rather than "click here".10 Every article should link to two to five related pages on your site, and your most visited pages should link to the pages you most want to rank. If you have written several pieces on one subject, gather them under one overview page that links to each of them and is linked from each of them. That cluster tells both readers and search engines what you know about.
15. Fix the broken links
Run a crawl with a tool such as Screaming Frog and look for two things: links to pages that return a 404 error, and redirect chains where one old address points to another old address. Fix the link, redirect the old address to the closest live page, or remove the link. It is basic hygiene, and it is the kind of work that is never urgent until a lost page turns out to have been the one with the backlinks.
16. Refresh the pages that used to bring traffic
In Search Console, sort your pages by clicks over the last 16 months and look for the ones that have fallen. Often the content is simply out of date: last year's prices, a tool that no longer exists, a law that has changed. Update the facts, add what readers now ask, tidy the structure, and only change the date shown on the page when the update is substantial. A page that already has links and history usually recovers faster than a new one can climb.
17. Add structured data where it still earns something
Structured data helps search engines understand a page, and for some types it still earns a richer result. Google's search gallery lists the types it supports, among them product snippets, merchant listings, review snippets and organisation details.11 For a shop, product markup with price and availability is the one to get right first. Do not add markup to chase features that no longer exist. Google limited HowTo and FAQ rich results in 2023, and since May 2026 no longer shows FAQ rich results at all.1213 Validate with the Rich Results Test, and make sure the markup matches what is visible on the page.
18. Target the questions behind the keyword
You will not rank for "accounting software" against companies with a hundred times your budget. You can rank for "accounting software for sole traders with a Shopify shop", and the person searching for it is much closer to buying. Write down the questions customers ask you on the phone and in emails, in their words, and give each of the important ones a section or a page. This is also how people now type into AI assistants: in full questions, with their situation attached.
19. Check how your pages score on Core Web Vitals
Google's Core Web Vitals measure loading, responsiveness and visual stability. The thresholds for a good experience are a Largest Contentful Paint within 2.5 seconds, an Interaction to Next Paint of 200 milliseconds or less, and a Cumulative Layout Shift of 0.1 or less, measured at the 75th percentile of real visits.14 Interaction to Next Paint replaced First Input Delay in March 2024, so any checklist still mentioning First Input Delay is out of date.15 Search Console shows which groups of pages fail. The usual causes are oversized hero images, fonts and scripts that block rendering, and adverts or banners that push the content down after it has loaded.
20. Earn links with something worth citing
Links from relevant, respected sites still matter. Buying them, or trading them in bulk, breaks Google's spam policies and can get a site demoted.16 The honest route is to make something other people want to reference: a calculator, a price index, a clear guide to a regulation, a survey of your own customers. Then tell the handful of people who write about your subject that it exists. One link from a trade publication your customers read is worth more than a hundred directory listings.
21. Give images names, alt text and the right size
Google uses the page around an image, its file name and its alt text to understand what it shows.17 Rename "IMG_0001.jpg" to something descriptive before upload, write alt text that describes the image for someone who cannot see it, and serve each image at the size it is displayed. The same work improves accessibility, image search and speed at once.
AI search: being the answer, not just a result
Google's AI Overviews, ChatGPT search and Copilot now answer many questions directly and cite a few sources inside the answer. Google states that there are no additional requirements or special optimisations needed to appear in AI Overviews and AI Mode, and that ordinary SEO best practice applies.18 What follows is that best practice, applied to being quoted. The full mechanics are in Beyond SEO: winning visibility in the age of AEO and GEO.
22. Answer the question in the first sentence
Use the question as the heading, and answer it in the first sentence under it, then add the detail, the exceptions and the example. A page that takes five paragraphs to reach the point gives an engine nothing to lift. The same structure wins featured snippets, which Google picks from pages that answer a query directly.19 Read your ten most visited pages and move the answer to the top of each section.
23. Show who wrote it and why they know
Google describes the qualities its systems look for as experience, expertise, authoritativeness and trustworthiness, and says that trust is the most important of them.20 Put a real author on every article, with a short line on why that person knows the subject and a link to a profile page. Name your company in full with its address on the contact page. Cite your sources. None of this is a trick, it is how readers decide whether to believe you too.
24. Publish one number nobody else has
Engines and journalists cite figures. If you have data nobody else has (average delivery times across your orders, the most common questions your support team gets, a small survey of your customers), publish it with the method in one paragraph. The research paper that named Generative Engine Optimisation found that adding citations, quotations and statistics to a page could raise its visibility in generated answers by up to 40%.21 This is a job for more than an afternoon, but the afternoon is enough to decide which number you will publish.
25. Be mentioned elsewhere, under one consistent name
Language models learn what a business is from what the rest of the web says about it. Use exactly the same company name, description and address on your site, your Google Business Profile, LinkedIn and the directories that matter in your trade, and link them together from your site with organisation markup. Then get mentioned where your customers already read: a guest article, a podcast, an answer in a trade forum. Twenty short, similar posts on your own blog do less than one page other people refer to.
26. Decide which AI crawlers may read your site
This one takes ten minutes and most businesses have never done it. OpenAI separates its crawlers: OAI-SearchBot decides whether you can appear in ChatGPT search answers, while GPTBot collects content that may be used to train its models, and you can allow one and block the other in robots.txt.22 Google offers a separate token, Google-Extended, to control whether content it crawls may be used to train and ground its Gemini models, without affecting Google Search.23 Open your robots.txt, check what it says, and make the choice on purpose rather than by default.
Email: the list you own
An email list is the one audience no platform can take away from you. It also comes with rules, so the first tactic here is the legal one.
27. Check that every signup asks properly
In the UK, marketing emails to individuals need their consent, with one exception: the "soft opt in" for existing customers, which covers contacts you collected during a sale or negotiation, only for your own similar products, and only if you gave them a simple way to refuse at the time and in every message.24 Look at every form that adds people to your list. Is there an unticked box or a clear statement of what they will receive? Is the unsubscribe link in every email? Your list is only worth something if you are allowed to write to it. The details of consent friendly capture design are in GDPR vs growth.
28. Clean the list, because the inbox providers now insist
Since February 2024 Google requires everyone sending to Gmail addresses to authenticate their domain with SPF or DKIM and to keep the spam rate reported in Postmaster Tools below 0.3%. Bulk senders must also offer one click unsubscribe.25 Sending to people who stopped reading long ago pushes that spam rate up and drags down delivery for everyone else on your list. Send one honest "do you still want these?" email to subscribers who have not clicked in six to twelve months, then remove those who do not respond.
29. Segment by what people did, not who they are
A new subscriber and a customer of five years should not get the same email. Split the list by behaviour: bought or not yet, bought once or several times, which category they bought from, which form they signed up through. Start with two segments, customers and not yet customers, and write the next campaign twice. It is more work per send and fewer sends that annoy.
30. Measure clicks, not opens
Apple's Mail Privacy Protection prevents senders from seeing whether a message was opened, and it is offered to everyone who reads mail in Apple Mail on an iPhone, iPad or Mac.26 Since then, open rates have been inflated and unreliable, and any tactic built on them (resending to "non openers", choosing a subject line by open rate) rests on a broken signal. Judge campaigns by clicks, replies and orders. If you resend a campaign, resend it to people who did not click, with a new subject line, once, and watch the unsubscribe rate.
31. Write subject lines that say what is inside
A good subject line is specific: "Your order ships Thursday", "Three changes to our delivery prices", "The 2027 catalogue is here". Keep the important words at the start, because a phone shows only the first few. Test one change at a time against a split of your list, judged on clicks, and test send times the same way rather than trusting a rule about Tuesday mornings.
32. Send from a person
"noreply@" tells the reader not to engage. A real name ("Sarah at Harbourside Tea") with an address that accepts replies feels like a person writing, and replies are a signal inbox providers read as engagement. Make sure someone actually reads those replies.
33. Build a welcome series instead of one welcome email
The moment someone subscribes is the moment they are most interested. One welcome email followed by silence until the next newsletter wastes it. A series of three to five emails over two or three weeks can introduce what you do, deliver what the signup promised, answer the most common objection and make one clear offer. Write it once and it works for every new subscriber.
34. Try plain text for messages that should feel personal
A heavily designed email looks like an advert and gets treated as one. For relationship messages (a thank you after the first order, a question to lapsed customers, a note from the founder), test a plain text email that reads like one person writing to another. Keep designed templates for catalogues and offers where pictures carry the message.
35. Give each email one job
Five links to five places leave the reader to decide what matters. Decide the single action each email is for, put that button near the top, repeat it at the end, and cut or demote everything else. Short paragraphs and a few bold phrases help people who skim, which is most people reading email on a phone.
Paid advertising: wasting less of the budget
Every pound spent on clicks that could never convert is a pound not spent on ones that could. These tactics are about cutting waste and giving the platforms better information.
36. Set up conversion tracking, with consent, before anything else
If you do not know which campaigns produce orders or leads, every other ad tactic is guesswork. Track the actions that matter (purchases, qualified leads, bookings) and check that the numbers in the ad platform roughly match the numbers in your shop or CRM. Do it with consent. Under the UK's Privacy and Electronic Communications Regulations, storing or reading information on a visitor's device needs consent unless an exception applies. Since 5 February 2026 the exceptions in Schedule A1 include analytics used purely to improve your own site, provided you tell visitors and let them object, but advertising pixels are not covered.27 The regulator is now the Information Commission (formerly the ICO), which replaced the Information Commissioner under sections 118 and 119 of the Data (Use and Access) Act 2025.28 Google's consent mode has a basic version, in which Google tags do not load at all until the visitor answers the banner, and an advanced version that sends cookieless pings before consent.29 For a UK site, which version you run is a risk decision to make with your lawyer. Server side tracking of people who declined can be used for counting, never for retargeting. The full setup is in event tracking and consent mode.
37. Add negative keywords every week
Open the search terms report for your search campaigns and read what people actually typed before they clicked. You will find job seekers, students, people looking for a free version, and competitors' brand names you never meant to bid on. Add the irrelevant ones as negative keywords and keep a shared list across campaigns. Fifteen minutes a week, and it is often the single largest saving in an account that has been left alone.
38. Send each campaign to a page that matches it
An advert for "same day flower delivery in Bristol" that lands on the home page makes the visitor search again. Build a landing page for each main campaign whose headline repeats the promise of the ad, with the offer, the proof and one action, and without the full navigation. It also raises the landing page experience part of Google's quality score, which lowers what you pay per click.
39. Test creative on a calendar, because it now does the targeting
The platforms increasingly choose the audience themselves. Meta's Advantage+ campaigns and Google's Performance Max take broad targeting and find people based on who responds to the ad. That means the advert itself has become the main lever you control. Keep at least one challenger running against your best ad at all times, change one thing per test (the opening line, the image, the offer), and retire the losers on a fixed day each month.
40. Retarget only the people who said yes
Showing ads to people who visited your site works because they already know you. It is also exactly the kind of tracking that needs consent. Build retargeting audiences only from visitors who accepted marketing cookies, and from customers who agreed to marketing, never from people who declined. Exclude recent buyers so you are not paying to show them what they just bought, and cap how often the same person sees the ad.
41. Schedule ads around when people buy
Look at conversions by hour of day and day of week over the last three months. If almost nobody buys between midnight and six, there is little point paying full price for clicks then. Lower bids at those hours rather than switching off completely, because the data for small accounts is noisy and a few late orders can be worth keeping.
42. Give automated bidding a clean signal before you hand it the wheel
Automated bidding strategies such as target cost per acquisition or target return on ad spend learn from the conversions you report. If you count newsletter signups and purchases as the same conversion, the algorithm will happily buy you newsletter signups. Before switching to automated bidding, make sure the primary conversion is the one that makes money, give it a value where you can, and leave the strategy alone for a couple of weeks after each change so it can learn.
43. Untangle the account structure
Accounts that have grown for years tend to have overlapping campaigns bidding on the same searches, duplicated audiences and budgets split so thin that no campaign gets enough data to learn. Write down what each campaign is for in one line. Merge the ones that do the same job, pause the ones nobody can explain, and make sure each remaining campaign has enough budget to collect conversions.
44. Work out what a customer is worth and what one costs
Cost per acquisition on its own tells you nothing about profit. The comparison that matters is customer lifetime value against customer acquisition cost. A simple version of lifetime value, for a business where customers come back to reorder, is:
Here is the average order value, is the average number of orders a customer places over the period you care about (two years is a sensible horizon for a small shop), and is the gross margin as a share of the order value. The reasoning behind it is in the Foundations on lifetime value. Customer acquisition cost is:
where is everything you spent on marketing in a period, including tools and agency fees, and is the number of new customers in that period, a plain ratio explained in the Foundations on ratios and rates.
Worked example: Harbourside Tea in Bristol
Harbourside Tea (harboursidetea.example) is an invented online tea shop in Bristol. Last quarter it spent £3,000 a month on advertising and won 75 new customers a month. Its average order is £32, its gross margin 45%, and its customers place 2.5 orders on average over two years.
So each new customer costs . The lifetime value uses , and :
So : on average, each new customer costs £40 and returns £36 of margin over two years. The business is buying growth at a loss. Splitting the same figures by campaign shows where. The brand search campaign spent £810 for 45 customers, who go on to place 3.2 orders on average. The broad shopping campaign spent £2,190 for 30 customers, who place 1.45 orders.
Brand search customers cost each and are worth in margin. Now the broad campaign:
Here is against of . The broad campaign costs £73 to win a customer worth £20.88. The fix is not to switch it off blindly (some brand searchers first discovered the shop through it), but to cut its bids sharply, test it on a narrower product range, and watch whether brand searches fall. It also shows the limit of the brand figure: some of those 45 customers would have found the shop anyway, so £18 flatters it. For a deeper look at bidding on value, see bid on lifetime value, not cost per click.
Five this afternoon, five next week
Forty four fixes are too many to do at once, and doing them all at once would make it impossible to tell which ones worked. Pick five. If you are not sure which, take conversion tracking with consent (36), the phone call to action (1), the form fields (4), the ten page titles (12) and the search terms report (37). Write down the date and what you changed. Next week, look at the numbers, keep what helped, and pick five more.
Growth hacking as a label has a mixed reputation, partly because of exactly the kind of invented figures the old version of this article carried. The useful part of it is the habit underneath: small changes, measured honestly, repeated every week. That habit, kept within the consent rules, is what privacy first growth hacking is about.