WooCommerce and garden centre
Ten years of seasonal advice, all of it still live
WordPress never throws anything away. A garden centre that has been blogging since before it had a shop is sitting on hundreds of posts, each written for one month of one year, every one of them still published at its own address and still perfectly readable. That archive is the richest material any garden centre owns and the most reliable way to get a confidently wrong answer, because nothing in it says which month it is now.
Why this pairing is its own job
The WooCommerce install guide covers the tag, and the garden centre guide covers what the assistant has to know. What follows is the part that belongs to neither.
- Advice to plant something is wrong for most of the year, and the archive contains advice for every month written in the present tense. An assistant matching on the topic rather than the timing will answer a November question out of an April post, and the customer will do what it says.
- Date archives, category archives and tag archives republish the same text at several addresses, so one enthusiastic post about spring bulbs can be matched several times over and drown out the page you actually maintain.
- Garden centre sites carry a consent banner more often than most, and consent plugins hold third party scripts until somebody accepts. Every visitor who ignores the banner sees no launcher at all, which looks exactly like a broken install.
- The cafe, the events, the Christmas grotto and the hours all sit on the same WordPress site as the shop, so the tag reaches them for free. That is the one thing this platform gives this trade that the hosted stores do not.
What changes about the install here
<script src="https://cdn.askably.xyz/w.js" data-key="pk_live_YOUR_KEY" defer></script>The tag is the same one on the WooCommerce guide. Everything below is about where it goes on a garden centre site specifically.
The consent banner decides whether the widget exists
Consent plugins classify scripts and hold everything outside the necessary category until a visitor agrees. Until then the loader never runs, so there is no launcher, no console message about an origin, and nothing to debug. The install is fine. It is simply not allowed to happen yet.
Whatever category you place it in is a decision you should be able to explain. Putting a chat assistant in the strictly necessary bucket to make it appear for everybody is a choice with consequences beyond this widget, and it should be made with whoever owns that decision rather than by whoever is pasting the snippet.
Test it the way a visitor meets it: a private window, banner untouched, then the same window with consent given. If the launcher only appears in the second case, you now know what your real coverage is.
Decide what the crawl is allowed to reach
It obeys your robots file, which makes robots the cheapest lever you have here. Date archives and tag archives add nothing but duplication, and posts older than a few seasons add advice you would not stand behind today.
Being blunt about this is better than being thorough. A crawl of two hundred well chosen pages produces sharper answers than a crawl of two thousand pages where nine tenths of them contradict each other about when to prune.
One snippet covers the whole site, which is what you want here
For a pure store, the interesting question is which pages to keep the widget off. For a garden centre it is the opposite: the events page, the cafe menu, the Christmas opening times and the school visit page are all questions somebody wants answered, and they all sit on the same site.
Leave it running everywhere and spend the effort on the material instead. The one place worth a conditional is the checkout, and on a garden centre site that is a small share of the traffic anyway.
Where the answerable material lives
| One seasonal calendar page that replaces the archive | Twelve short sections, one per month, saying what is worth doing and what you have in. Written once and edited each year rather than added to each year. This is the page you want matched instead of a post from four seasons ago, and it only wins if it exists. |
|---|---|
| Care sheets by plant group, as pages rather than posts | Watering, feeding, soil, aspect, hardiness, eventual size and pruning window, written by group. Group level survives stock changes and covers far more questions per document than a per variety page, and pages sit outside the dated archive so they read as current. |
| Visiting information, kept in one place | Hours through the year, dogs, the cafe, parking, accessibility across gravel and outdoor tills. On a WordPress site this is usually scattered across three pages and a post from 2019, and consolidating it is half an hour that pays back every weekend. |
| The guarantee and delivery terms, current version only | One page each, replacing rather than superseding whatever came before. Old terms left published are the same problem as old advice: they are still true looking, still crawlable, and still capable of being quoted back at you. |
The first thing to get right
Fold your ten most read advice posts into one calendar page you own, then block the date and tag archives from the crawl.
It swaps a pile of contradictory dated material for a single document you can correct, and it removes the duplicate addresses that let the same old post win a match repeatedly. Both halves matter: writing the page without cutting the archive leaves the archive competing with it.
The failure that belongs to this combination
It quotes a post from 2017 and sounds like it works there
This is the failure that is hard to catch, because the answer is good. It is in your voice, it is specific to your region, it names plants you actually sell, and it carries a citation pointing at a real page on your own site.
The only thing wrong with it is the month. Somebody reads it in November and does a job that needed doing in April, and when the plant fails they blame the plant and then the shop.
The citation is where you spot it, because it points at an address with a year in it. Read the citations on a set of seasonal test questions before you trust the answers, and if a dated post is winning, the fix is to remove it from the crawl rather than to argue with it.
Before you go live
- The origin the page is served from has to be on the allowlist for that assistant, or nothing renders and the browser console says which origin was refused. An apex domain and its www are two different origins to a browser, so list both, along with any staging or preview host you want it to work on.
- Open the site as a visitor would, on the pages a garden centre visitor actually lands on, and ask it something only your own material could answer. A widget that renders is not the same as a widget that has read anything.
Questions
- Can we not just crawl the whole site?
- You can, and on a young site it is the right call. On a site with a decade of dated advice it is the thing that makes the assistant unreliable, because the archive is larger than the material you maintain and it will win on volume. Narrow the crawl and grow it back.
- The banner is stopping it from appearing. Is that a fault?
- No, that is the consent plugin doing its job. The loader is being held until a visitor accepts, so nothing renders and nothing is logged. Which category it belongs in is a decision for whoever owns your consent policy, and it should be made explicitly rather than by whoever installs the snippet.
- The shop is a small part of our site. Does that change anything?
- It changes what you index first. The events, the cafe, the hours and the seasonal jobs are what people ask about, and they are not product pages. Index those before the catalogue, and treat the shop as one more section of the site rather than as the point of it.
Keep reading
- Installing on WooCommerceHook the script into wp_footer so it survives theme updates, keep it out of the parent theme, and stop a caching plugin from mangling it.
- Everything for a garden centreSeasonality drives every question a garden centre gets. Plant care answers well, toxicity and site suitability need a person.
- A bike shop on WooCommerce, where the sizes and the workshop are both invisibleFrame sizes are variation data rather than page text, and the workshop queue is not a product at all. Both are why a bike shop store answers badly by default.
- Excluding a regulated range on a store whose addresses do not distinguish itProduct permalinks carry no category, so the pharmacy range cannot be excluded by path. And on this platform the checkout is yours to put a widget on.
- A garden centre on WordPress, where the archive is the problemWordPress keeps every seasonal post at its own address forever, so a crawl reads April's advice in November. In this trade that is not stale, it is wrong.
- A garden centre whose questions are mostly not about the shopHours, dogs, the cafe and the compost delivery area are the real questions, and none of them live in the shop. Where to put them instead.
Try it on your own material
Upload a document or point it at your site, paste one line of HTML, then ask it something only your business could answer.