See Personyze at work, one feature at a time

• 3:07Now playing01A whole campaign: home-page bannerOne campaign, start to finish, on a live store: shoppers who looked at footwear come back to the home page and find a banner written for them. The audience in two rules, the banner and its words, a new slot picked on your own page, and the store before and after.
• 1:40Now playing02Installing the tracking codeNothing works until the tracking code is on your site. One snippet for every domain on the account, in the head of each page — copied, sent to whoever manages your site, added through Google Tag Manager or by your platform's guide — and the tracker firing a few seconds after a visit.
• 2:09Now playing03Setting up the product catalogThe products your recommendations choose from. Guided setup asks where they are — your store, a feed, your own site, a file or SFTP — matches your columns to Personyze's fields, checks the products the way a shopper will see them, and shows what the engine has to work with.
• 1:35Now playing04Setting up product trackingRecommendations learn from what shoppers do with your products: what they view, add to cart and buy. Three ways in — Google Tag Manager, reading it from your pages, or a line of code — a live check that ticks as each event arrives, and past orders uploaded so the engine starts from what sold together.
• 1:55Now playing05Setting up the content catalogThe articles your content recommendations choose from. Guided setup asks where they are — an RSS or Atom feed, your own site, a file or SFTP — matches your columns, checks the articles the way a reader will see them, and lists them, ready to recommend.
• 1:02Now playing06Setting up article trackingContent recommendations learn from what readers open and read. Read it from your page addresses with no code, or add one line on every article page; a live check ticks as a read arrives, and past reads can be uploaded so recommendations know what is popular from day one.
• 2:28Now playing07Product recommendations, automaticA product recommendations campaign checks your catalog and shopper events first, then the widget: a design to start from and Automatic — one widget for every page that decides by itself, falling back step by step so it is never empty — placed in a slot on your pages and published to staging.
• 2:31Now playing08One recommendation: bought togetherOne specific recommendation: what other shoppers bought together with the product on the page. Cross-Sells among the algorithms for a product page, how many to show, a fallback for products with no orders yet, skipping what is already in the cart, a title, and a new slot under the product.
• 1:15Now playing09Product recommendation resultsWhat a recommendation campaign earns: the purchases that came from a recommendation, their revenue and share of all sales, every step from shown to bought against the period before, and each product's own numbers — which sell when shown, and which are looked at and skipped.
• 1:14Now playing10QA and the recommendation SimulatorBefore a recommendation goes live: QA sums up who it reaches, whether it runs and whether another campaign is in its way; your own product page opens with a bar showing what showed and which algorithm filled it; and the Simulator shows what it would pick for a product, or for a real visitor.
• 2:36Now playing11Content recommendations, automaticA content recommendations campaign checks your content catalog and reader events first, then the widget: article rows from a design, and Automatic — one widget for every page that decides by itself what fits, falling back step by step — placed under every article and published to staging.
• 2:29Now playing12One recommendation: readers also readOne specific content recommendation: what readers of this article went on to read. The algorithms for an article page, how many to show, a fallback of similar articles for a new post, skipping what they have already read, a title, and a new slot right after the article text.
• 4:17Now playing13Recommendation filters and fallbacksOne recommendation widget built properly: the algorithm families on a product page, a filter whose value comes from the product being viewed — same brand — a fallback algorithm for when the first runs out, and a separate widget title for each algorithm, then both outcomes on a live shop.
• 0:48Now playing14Show visitors what they searched forWhat a visitor searches for is the clearest thing they will ever tell you. Personyze keeps each visitor's last search on their profile - from your site search, or the keyword of the search ad that brought them - so you can target by it with the Last search rule, or recommend exactly what they searched for on any page with the Matches Their Search algorithm.
• 4:16Now playing15Content recommendation filters and fallbacksOne content recommendation widget built properly on a B2B blog: the algorithm families on an article page, co-reads, boosting promoted and new articles, a filter whose value comes from the article being read — same topic — a fallback for when co-reads run out, and a separate title for each algorithm, then both outcomes on the live blog.
• 1:36Now playing16Setting up category recommendationsAdd a Category Recommender — a shop-by-category block that picks categories for each visitor — and place it on your product page.
• 3:09Now playing17Trending and bought-together setsOn a real store’s data: which products are gaining or losing momentum, and the product pairs the recommendation engine learned from shoppers — opened on a Sony Bravia 75-inch TV.
• 1:27Now playing18Product recommendations in emailPut personalized product recommendations into any email platform. Pick an email template, choose what to recommend, and paste one block of HTML — every open renders products for that recipient.
• 2:18Now playing19Setting up an A/B testAn A/B test shows different versions to different visitors and tells you which works better. Everyone who visits the store takes part; Group A gets free shipping, Group B 20% off the first order, and the Original sees neither; the winner is picked on purchases, with auto-pick, and the test goes to staging.
• 1:55Now playing20Setting up campaign targetingDecide who sees your campaign. Describe an audience in plain words and let AI write the rules, or build them yourself — then check the result against real traffic.
• 2:04Now playing21Advanced targeting rulesGo deeper into targeting: let AI write rules from a sentence, explore every rule in the library, reuse saved audience templates, and combine rules with AND, OR, XOR and groups.
• 1:43Now playing22Auto-pick winner & AI optimizerLet Personyze decide your A/B test for you — declare and deploy the winner automatically, or keep optimizing the traffic split with AI every night.
• 4:26Now playing23A/B test campaign, end to endA whole A/B test, start to finish: who takes part, what each version does — a page edited in place, personalized to the visitor, plus a form — the traffic split with a control group, the goal it is judged on, and the report that follows.
• 1:01Now playing24Reading A/B test resultsAn A/B test reports on its Performance step, and it opens on the verdict: which version to keep, how sure Personyze is, and where the answer differs by audience - here, on phones the free-shipping version leads. Below it, how traffic was split and each variation's visitors and conversion rate; then every goal, each variation against the baseline, how far ahead or behind it is, how sure Personyze is and the range the real difference lies in. The conversion rate over time shows when the result became clear, decision criteria set how much each goal counts when several pull in different directions, and the breakdown by audience shows who responded to which version.
• 1:38Now playing25A/B winners for each audienceAn A/B test has one winner overall - but not always for everyone. A/B tests by audience slices a test by the audiences Personyze discovered in your data, and marks a Clear winner only when it holds up: corrected for how many audiences were checked, ahead in both halves of the test, with enough data. Tick the audience and apply: each audience sees the version that won it, the campaign is saved as a new version, and Undo puts the test back.
• 1:10Now playing26Setting up conversion goalsSet up a conversion goal: choose what triggers it, count it the way you need, decide where it is used, and give it a value.
• 1:33Now playing27Campaign goalsMeasure a campaign by the goals that matter to it: account-wide goals by default, plus a goal dedicated to this campaign that uses no account goal slot.
• 1:23Now playing28Automation rulesLet Personyze watch your numbers: a rule that checks every campaign, emails you when one is being dismissed more than it is used, and can post to Slack, call a webhook, or pause and publish campaigns.
• 2:17Now playing29Setting up AI Search and ChatAn assistant that answers your visitors in your own words, from your own site. What it answers from, what it may know about the visitor, who takes over when someone asks for a person, who meets it and where it sits — a chat bubble, an ask box or a finder — then a shopper asking on the store.
• 2:48Now playing30Meeting booking on your siteVisitors book time with your team without leaving your site. Hosts, their hours and what can be booked are set once for the account; a campaign decides who is offered it and where the booker sits — then a visitor picks one of the host's real free times, in their own timezone.
• 2:46Now playing31Chat inbox and human handoverEvery conversation your AI assistant had, sorted by how it ended — and a visitor handed to a person live: the operator screen picks it up, reads what was already said, and the reply lands in the widget the visitor is still using.
• 1:53Now playing32AI Product FinderThe AI Product Finder is a bar at the foot of your page where a shopper describes what they need in their own words. Your AI assistant reads the need against your product catalogue and answers with a results page: a short summary, then the products it picked, each with the reason it fits. Set the bar's words and example chips, preview every state without spending AI credit, shape the results page, and place it as a floating bar.
• 1:56Now playing33AI Content FinderThe AI Content Finder is a bar at the foot of your page where a reader asks what they want to learn. Your AI assistant reads the question against your articles and pages and answers with a reading list: a short summary, then the articles to read, each with the reason it fits, and a follow-up chat for more questions.
• 2:13Now playing34Setting up a personalization campaignBuild one campaign from start to finish: target people by the pages they viewed and by your CRM list, give them an embedded banner and a popup lead form, and publish.
• 2:36Now playing35Adding personalization actionsThe Content step holds what visitors actually see. Add an action, start from a template, edit it on the preview, and decide where, when and how often it appears.
• 1:15Now playing36Adding a placeholder to your pageCreate a placeholder — a named slot on your page where campaigns put in-page content — by clicking the spot on your live page.
• 1:33Now playing37Designing a popup (Canvas Builder)Design a popup or banner from ready-made blocks, with no code — edit text on the canvas, add elements like a countdown, and style the whole container.
• 2:02Now playing38Editing page content with AIChange your own site's pages — text, images and styles — right on the page, and let AI rewrite the copy for you. Your site itself is never modified.
• 1:48Now playing39Preview and QA before launchCheck a campaign on your own site before launch, see whether it shows to you, and find out exactly why when it does not.
• 2:06Now playing40QA simulator: test as any visitorOpen your own site with the QA simulator beside it and see, rule by rule, whether a campaign shows for the visitor you are pretending to be. Ask the AI assistant why it does not, see everything campaigns put on the page, and change the visitor to watch the answer change.
• 1:17Now playing41Building a lead capture formBuild a lead capture form: start from a template, set the fields and what each one saves into the visitor profile, choose how it appears, and decide where submissions go.
• 1:51Now playing42Social proof and countdown timersAdd urgency to your product pages: a social proof widget that only shows numbers worth showing, and a countdown timer — to a fixed date, or personal to each visitor.
• 2:29Now playing43Reviews: show and collectA reviews campaign does two jobs: it shows each visitor the reviews that matter to them, and it collects new ones on your site. Bring reviews in from 17 review platforms, a file, or by hand; show a product’s own reviews on its page; add a Review Form under them; and approve each new review before it goes live.
• 0:37Now playing44Reviews widget resultsA reviews campaign's report shows how the reviews you display work: how many visitors saw them, how many engaged, and how many of those visits converted, over time and against the period before. Then the visits it ran in beside the rest of your site - here, visits with reviews converted at 7.9% and the rest at 5.4%, and earned more per visit. To measure what the reviews themselves add, run the campaign as an A/B test with a group that sees none.
• 2:15Now playing45Feedback and NPS on your siteA Feedback & NPS campaign asks your visitors one question - how likely they are to recommend you, whether a page helped, stars or faces - with an optional comment. Every answer is private and lands on the visitor profile (NPS score and group) for targeting. The campaign dashboard shows the NPS over time with promoters, passives and detractors, and Feedback Results shows the score on your site once enough visitors have answered.
• 2:28Now playing46On-site surveys and resultsA Surveys campaign asks a few questions - a poll, an exit survey, a post-purchase survey - one at a time with a progress bar. Build the questions in the survey itself: stars, one or several choices, a 0-10 scale or a written answer; each answer lands on the visitor profile for targeting. The Survey dashboard shows respondents, response rate, how far visitors got, and the results per question. Survey Results then shares the totals on your site, only for the questions you make public and only once enough visitors have answered.
• 2:09Now playing47Surveys: logic, offers and sharingA Personyze survey knows who is answering. Here, an exit survey on a pricing page asks what is holding the visitor back, and each reason gets its own next step: a budget question, then a free proof-of-concept offer, then only the contact details the team needs. Questions are asked or skipped by plain-language rules, any text can include an earlier answer or a visitor detail, and the answers are saved on the visitor profile for targeting. The answers go to email, a webhook, Zapier, Slack or HubSpot - for every answer or only some. The same survey can have a page of its own, free on the Personyze address, with a QR code and a text-message link, and it can go in an email. The report shows the funnel from shown to answered, where people stop, where they answered, every question with the rule that decided who was asked, and an AI summary of the written answers.
• 2:40Now playing48Promotions: the right banner per visitorA Promotions campaign puts your promotions on the site and lets an algorithm choose which one each visitor sees. Its wizard starts with your promotions list and a five-week calendar by page category. In Content, choose image banners, text promotions or coupons and a design - here a banner across the category page. Then the algorithm: Matches this category shows the promotion for the category being browsed, others go by recent categories, search, location or popularity, and each says whether your data supports it; with nothing for a category it falls back to a site-wide promotion. Learn from results shows the better earners more often, and a holdout can keep a share of visitors promotion-free to measure the lift. Set the look, place it in the category banner slot, read what the visitor will actually get, and publish: each category page gets its own banner. The Promotions Overview and Report show views, clicks, codes, purchases and revenue per promotion.
• 2:21Now playing49Display rules and popup triggersEvery setting that decides how an action appears: inline or as a popup, its position and behaviour, the triggers that open it, frequency caps, and which pages it may run on.
• 2:01Now playing50Frequency capping a campaignLimit how often a whole campaign shows to one visitor: once ever, once per session, until they click or close it, a few times then a break — or your own conditions.
• 3:29Now playing51Campaign conflicts and overlapsEvery way to settle campaigns that compete for the same visitor: rules on other campaigns and actions, audience persistence, audience groups with priorities, placeholder order and timing between actions.
• 2:21Now playing52Building a hosted landing pageBuild a landing page that Personyze hosts for you — on our address or your own domain — personalize it per visitor with variables and versions, and A/B test the page itself.
• 1:00Now playing53Capturing visitor attributesCapture data about every visitor to personalize on: session values, profile fields and cross-device identifiers — and a new container that reads the campaign from the link.
• 2:27Now playing54CRM targeting and personalizationPersonalize for people you already know. Recognise members from a link parameter, upload your CRM list, target that list, and print each member’s first name into what they see.
• 3:00Now playing55Salesforce CRM, both waysPersonyze personalizes your website with your Salesforce CRM data, and sends what the website learns back into Salesforce. Your Salesforce admin connects it once, in about ten minutes: an External Client App in your own org, then your My Domain address, a Consumer Key and a Consumer Secret in Personyze. Choose the fields that arrive on the visitor - from their Contact or Lead, their Account and the Account's open opportunity, custom fields included; Personyze looks a known visitor up once per visit. Anonymous visitors from target accounts are matched to their Salesforce Account by the website domain that company identification finds. Forms and meeting bookings update the matching Contact or Lead or create a new Lead; after each visit, the profile sync writes audiences, interests and visits to the person's record when something changed, and can log the visit as a Task; Salesforce Campaigns become audiences, refreshed every hour. Then a campaign targets an open deal's stage and names the account in the home page headline. Salesforce is a trademark of Salesforce, Inc.
• 1:25Now playing56Audience Discovery: find your audiencesAudience Discovery reads your own visitor data every week and finds the groups that reach your goals far more - or far less - often than average. Each discovered audience is a short, readable rule, with how much more often its people convert and the evidence behind it, checked on visitors it was not fitted on. Choose what counts as success and which data the rules may use, then use any audience in a campaign, turn it into a list, or see how every campaign and A/B test performs for each audience.
• 1:56Now playing57Audience discoveryPersonyze reads your traffic on its own and reports the groups that behave differently — who comes back more, who engages more, who performs worst — with the evidence behind each one.
• 1:34Now playing58Connecting your integrationsConnect Personyze to your stack — analytics, advertising, CRM and marketing automation, lead intelligence and developer tools — with Google Analytics 4 set up as the worked example.
• 1:49Now playing59Traffic and campaign reportingThe two reporting screens: your whole site — visits, quality, goals and referrers — and what your campaigns did, from impressions and clicks to conversions per goal and per action.
• 2:09Now playing60Live visits in real timeWatch sessions as they happen: who is on the site, what they are doing, which campaigns are matching them — and open any session to see every page and event in order.
• 1:16Now playing61Scheduling reports by emailGet your Personyze statistics by email on a timetable — for the whole site or one campaign, weekly, daily or monthly, to everyone who needs them.
• 4:40Now playing62Setting up an ABM campaignAn account-based campaign end to end on a B2B site: an audience defined by the company behind the visit — its size and its CRM stage — then the home page call to action rewritten with the on-page editor and personalized with the company name, and a block placed on the pricing page for large accounts.
• 2:18Now playing63ABM landing page by accountA B2B vendor’s website speaks to one target account at a time. A visitor from a target bank arrives, Albacross names the company behind the visit, and the same landing page becomes a page made for that bank — its logo and name, a line written for it, the case study that fits, more stories from its own country first, and an exit offer in its name. One campaign, four actions, and every bank-specific word comes from the account record.
• 2:01Now playing64Welcome bar for newsletter visitorsVisitors who arrive from your newsletter link get a welcome of their own: an audience built on the landing page’s UTM source, an announcement bar written for them in the slot your site already has — and the shop with and without the newsletter link.
• 2:56Now playing65Report form for engaged readersOn a B2B blog, readers who open a second article are offered the report — a lead form placed after the article text, filled in and sent on the live site, with the reader’s name and email landing on their profile.
• 1:24Now playing66Discount code, tested with a holdoutGet a discount code into the right hands and prove it earned something: build it as a Program, test two coupon designs against each other, and hold back a slice who see nothing.
• 1:37Now playing67Abandoned cart recoveryWin back abandoned carts: email people their own cart, hold back a slice to prove it worked, and welcome the ones who click through with a reason to finish the order.

Nothing matches both filters. Show all 67 videos

More walkthroughs are on the way.