=== Factorize Simulate ===
Contributors: webyoungstown
Tags: simulation, customer research, woocommerce, pricing, ai
Requires at least: 6.2
Tested up to: 7.1
Stable tag: 1.7.2
Requires PHP: 7.4
License: Factorize License (proprietary)
License URI: https://factorize.webyoungstown.com/license/

Ask a question about your store and watch a crowd of simulated customers react, before you act.

== Description ==

Type a question such as "How will customers react if we raise prices 8%?". Factorize Simulate sends your question and the material you choose to Factorize's servers, where a crowd of made-up customers reads about your situation and talks to each other on pretend social sites. You then read everything they said, in your own WordPress admin.

It is a rehearsal for finding problems and wording to fix, not a sales forecast. The people are simulated.

= What you get =

* **A verdict** with how sure it is, why, and the caveats.
* **Sentiment** over time as drawn charts, by place, with the turning points and the posts behind them.
* **The crowd**: groups, the most influential people, and every person with their persona, filterable and searchable.
* **Posts**: every post and reply, filtered by place, round, feeling and theme, with highlights.
* **Themes**: objections, selling points, questions people asked, and topics.
* **A network** of who talked to whom, interactive, with a list you can use from the keyboard.
* **A graph** of what the people learned from your material, interactive, with a list you can use from the keyboard.
* **A written report**, ready to print or save as PDF.
* **Chat** with the report or any one person, and interviews of a group or of everyone.
* **A what-if**: change one thing (a competitor's sale at round 5) and compare the two runs side by side.
* **Cost under control**: you see the most a run can cost before you press Start. It stops at that limit, and unused tokens go back to you.

= The material you send =

You choose: paste text, upload a PDF, TXT or MD file, or switch on any of the store inputs under "What else the people read": products and prices, categories, your policy and About pages, reviews, order patterns, Store Stats, promos and auctions, the design brief. Each shows its size, a Preview and what it adds to the cost, and none contains a name, email, address or phone number (order counts by spend band, repeat rate, average order, busiest days, order counts by place, review words and up to ten review excerpts with personal details taken out). A "What will be sent" box shows the exact text before anything leaves your site.

= Tokens =

Runs on the cloud model use tokens: 1,000 tokens equal one dollar of what the model costs, with nothing added on top. Your first simulation is free. Packs are bought on the Tokens tab with Stripe's secure form. Simulations on Factorize's own computer are free for smaller sizes.

= Where results are kept =

Every finished result is saved in two tables in your WordPress database (`fzsim_runs` and `fzsim_parts`), so your history survives Factorize's own keep-time. You can turn this off in Settings and delete any simulation. Uninstalling the plugin removes both tables.

== External services ==

This plugin cannot work without the services below. It sends data only when you press Start, chat, interview, share or buy.

**1. The Factorize server** (https://factorize.webyoungstown.com, run by WebYoungstown, the maker of this plugin). Every request is signed with your site's Factorize key.

* What is sent: your question, your scenario, the material you chose to include (pasted or uploaded text, PDF files, and the store-numbers summary if you ticked it, exactly as shown in the "What will be sent" box), the size and settings of the run, your chat and interview messages, this site's address, and the Factorize key's site identity. For buying tokens: the pack you chose and a return address on your site. Nothing else from your site is sent; no names, emails, addresses or phone numbers of customers are included in the store-numbers summary.
* When: when you press Start, run a what-if, send a chat message or interview, open a simulation or the Tokens tab (to read the balance and the result), press View in MiroFish, or buy tokens.
* Terms and privacy: https://factorize.webyoungstown.com/license/ and https://factorize.webyoungstown.com/privacy/

**2. The cloud model provider.** Simulations on the cloud model are run by an AI model provider that Factorize chooses (at the time of writing DeepSeek, https://www.deepseek.com/). Your question, scenario and material, and the posts the simulated people write, are sent to the provider by Factorize's server to produce the simulation. Your site never talks to the provider directly. Runs on "our own computer" stay on Factorize's own hardware and are not sent to a provider.

**3. Stripe** (https://stripe.com), only when you buy tokens. On the Tokens tab, this plugin loads Stripe's script from https://js.stripe.com/v3/ and shows Stripe's secure payment form in a window. You type your card details into Stripe's form; they never reach your site or Factorize's server. Stripe receives your payment details and the amount. js.stripe.com is loaded on the Tokens tab only, never anywhere else on your site. Stripe's privacy policy: https://stripe.com/privacy

**4. The MiroFish viewer.** "View in MiroFish" opens an interactive view of a simulation on Factorize's viewer address, in a private link that lasts the number of days you set (7 by default). Anyone who has the link can see that simulation until it expires.

== Installation ==

1. Upload the `factorize-simulate` folder to `/wp-content/plugins/` and activate it.
2. Open Factorize → Simulate → Setup guide, and add your Factorize key in the License panel on the Factorize screen if you have not yet. A free key includes your first simulation.
3. On New simulation, type a question, read the "What will be sent" box, and press Start.

== Frequently Asked Questions ==

= Does it send my customers' names or emails? =

No. The store inputs build counts, bands and word lists, never a name, email, address or phone number, and review excerpts have those taken out. You can read the exact text before it is sent.

= What if I run out of tokens? =

Nothing is started, and nothing is charged. The message says how many tokens the run needs. Choose a smaller size, or buy a pack on the Tokens tab.

= Can a run cost more than the figure I saw? =

No. The most it can cost is reserved before it starts and it stops at that limit. Tokens it did not use go back to your balance.

= Do I need WooCommerce? =

No. Without it, you paste or upload your own material. With it, you can add a summary of your store.

== Changelog ==

= 1.7.0 =
* Design studio: pick up to three pages of your own store and a device (Computer 1280 px or Phone 390 px). The page opens inside wp-admin; hover and click any part and pin a comment on it: change this, add something here, remove this, move this, or what do people think? Up to 12 pins. Browse mode lets you open a menu or a product first.
* The simulated people now read the measured page: its reading order, the first screen, sizes, positions and colours, with each of your changes put in place, and every person answers for every pin.
* The Review step shows each pin with a plain "where" line and, folded, exactly what the people will read.
* The cost is shown step by step before you start, and after the run next to what each step really used.
* Results: a Pins card under At a glance with a wireframe of each page, numbered coloured dots, and a row per pin with its certainty, the range, who noticed it and the reasons. With many pins a pin is only called clear when it still holds up after allowing for the number of pins. Downloads: pin answers (spreadsheet) and wireframes (picture file).
* Visits made while looking at a page in the studio are not counted by the Factorize plugins that count visits.
* The older way, describing a design in words, is still there, folded.

= 1.6.0 =
* Overview: a short "At a glance" card above the verdict. The main finding in one sentence, a certainty as a big number with a line of words, the yes/no result when there is one, and up to three insights (the best-liked point, the strongest objection, the biggest difference between groups or designs when it is bigger than the noise).
* Certainty from statistics: the chance that the headline is right, worked out from the people's answers. Yes/no, design and feature tests use an exact Beta calculation; mood runs and paired design outcomes use a seeded bootstrap over people. Under 60% is "a coin toss", 60 to 80 "leaning", 80 to 95 "likely", 95 and over "very likely". It says how sure the simulated crowd is, not how real customers will behave.
* Ask everyone a yes or no question, in every kind of simulation: a question, two answer names (Yes/No, Buy/Don't buy, Use/Don't use, or your own) and the situation. The results get a Decision card (the share with its range, the certainty, the reasons on each side with quotes, by group) and a download of each person's answer. Needs engine 1.4.0.
* New simulation: "What else the people read" replaces the single store-numbers box. Products and prices, categories, store policies, the About page, reviews, order patterns, Store Stats, current promos and auctions, the design brief and your own text each have a switch, a description, their size in characters, a Preview of exactly what is sent and, on the cloud model, what they add to the token cost (derived from the server's estimate). Nothing is shown about tokens on the free computer or with an unlimited key. Everything is scrubbed; policies, about, products and categories start on.
* View in MiroFish opens the address the Factorize server gives, which now works from anywhere. If the server only knows an address on its own machine, the page says so instead of opening a dead link.


= 1.5.0 =
* New simulation: four kinds to pick from, Design, Feature, Product and General. Each has its own boxes, and "Also include" adds the others' (products, a feature, the design, files).
* Design test: press "Capture my site's design" and the plugin writes down your own pages (menu, reading order, clicks to a product, the cart and checkout, the phone view, theme, store facts) as text you can edit. Add one to three other designs in words, give the people one to five tasks, and every person tries every design, then says which they prefer. Screenshots are kept for you.
* New tabs: Design test (who wins and by how much, design by design, against the current design, what got in the way, task by task, by group) and Feature (would they use it, why, what would change their mind).
* Assurance replaces the single confidence number: a score out of 100 with five bars (sample, precision, agreement, stability, coverage) and a line on which part holds it back.
* Downloads for the test results, the walk-throughs and the choices. Needs engine 1.3.0 and a Factorize server that passes the kind of simulation on.

= 1.4.0 =
* New simulation: the size is an effort slider with four stops (Quick look, Standard, Thorough, Custom). One line under it says how many people and rounds, how long it takes and about how many tokens, and another says how big a difference it can tell apart (people, not posts, are what count).
* Overview: a "How solid is this?" card right under the verdict. Sample size, mood share with its range, average mood, how much the people agree, whether the mood changed over the run, who took part, and whether a repeat run agreed. Plain words, "bigger than the noise" or "within the noise", worked out from the result already on your site.
* Products say "too few mentions to tell" under 8 mentions, otherwise a mood with its range. Groups show how many people were scored and whether they differ by more than the noise.
* The Overview is shorter and in a better order: verdict, how solid, the report, the mood chart, themes, recommendations, downloads. A what-if says what it is a what-if of and what changed. The Graph tab names the kinds of things the graph found.
* With engine 1.2.2 every group (such as "Price-Sensitive") appears in the product by group grid. Older results keep working.
* "Look for runs on the server" now says why when the server could not be asked.
* The Factorize dashboard card no longer runs a database query on every admin page.

= 1.3.0 =
* Results: a Network tab, "Who talked to whom": people as dots (bigger = more active, colour = mood), lines for replies, quotes, likes and follows. Point at a dot to light up its neighbours, choose one to open that person, filter by place, zoom or fit, and "Show as a list" for the keyboard and screen readers.
* The mood chart is interactive (a small one on the Overview, full size on Sentiment): point at or tap a round to see each place's value and that round's most reacted-to post; choose a name in the key to hide a line.
* Products: a grid of products by group, coloured by mood. Point at a square for the posts and mood; choose it to read those posts.
* Needs nothing new from the server: the charts are drawn from the result already on your site.

= 1.2.0 =
* Product picker: searches the whole catalogue by name, SKU or ID (in stock or not), shows the true total and pages with Show more products.
* Results: a Products tab (how each picked product was received, with its own quotes), a Timeline tab (round by round), replies under posts, everything each person did, report contents and a Markdown download, a Download card, time by step, a readable setup summary; chat and interview answers stay after a reload.
* Needs the Simulate engine 1.2.0 for the Products tab and the right run dates.
* Factorize dashboard: Find a module sits right above the module cards; the 7-day order and sales tiles show the last 7 days with their dates.

= 1.1.1 =
* Waits are honest: a range (for example "about 4 to 24 minutes" in the cloud, "about 1.1 to 2 hours" on our own computer) on the size cards, in the cost line and in the Tokens table.
* Start works again when the hidden Custom boxes were empty; Enter in a text box no longer starts a run.

= 1.1.0 =
* Pick products as material: search your catalogue by name or SKU, pick up to 10; the people read each one's price, details, rating, a few reviews and units sold in the last 30 days. Works without Factorize Core.
* "Simulate this product" on the Products list.
* More questions (up to 5): answered after the run from the people's posts, with quotes. New "Your questions" section on the overview.
* Question ideas: ready questions for a price change, new product, policy change, ad test, new service, comparing two products, a launch, and discount depth.
* "Products in this run" on the overview.

= 1.0.0 =
* First release: New simulation, Simulations (overview, sentiment, crowd, posts, themes, graph, report, chat, run log and cost), Compare, Tokens (Stripe) and Settings.
