30 Ways To Scrape A Supermarket Website

Lets list and look at NZ supermarket comparison and shopping list tools, happy for you to tell me there are many missing, but I’m going to call this the OG 25…

PricePulse · Pricer · Grosave · Grocer.nz · CheckMyTrolley · ButterUp · Shoparound NZ · Trolly · ShopIt · PriceCrunch · GrocerCheck · Bargly · Tallo · T-Shop · ShelfScout · Cartsy (coming soon, apparently) · Basket NZ · Baskt · Kete · Is It Worth the Trip? · NZGrocerySaver · TrolleyMate · KiwiPrice · Appetise and Basket Brain

(This is a list by the way, not a hands on ranking of 25 apps.)

Each has a differentiated pitch, Some start with meal planning, others live in Home Assistant. Others on the list say coming soon.

So, 25 names.

Different stages, different jobs (but, really?)

A whole lot of the same groceries.

For a country of 5 Million, where apparently, butter needs a SaaS selector tool.

Some of these are useful

PricePulse compares prices while you shop on the supermarket websites, then lets you move the basket. Pricer totals your list at different stores. Grosave gives you price history and unit prices.

Is It Worth the Trip? counts petrol and your time.

I like that.

Saving $4 on cheese loses some appeal after a second car park and twenty minutes of your Saturday.

These differences do matter.

A person shopping online wants something different from someone scanning a barcode in the aisle. Getting either of them to actually use the thing is work too.

But read enough of the descriptions and they start merging, and become the same beast with a different face.

Lists, specials, history, cheapest basket.

Add an ‘AI meal planner’ and we can have the same conversation about mince in a chat window.

I would struggle to pick the next app from its feature list.

Show me how often its prices are wrong.

Butter problem

I'd guess AI helped build a fair few of these. I haven't checked how they were built.

Some look useful.

Others read like the same idea twenty times.

I can see the appeal.

Groceries are expensive, the problem is familiar, and you can get a first version working quickly.

A useful side project still need someone to use it, and somebody to pay.

Roc Wong built Grocer.nz as a side project in 2021, around a day job. Thousands of daily users. Someone offered him $10,000 for it. He turned it down.

He was right to. But that’s what the market thought a scraper, a matching table and a few thousand loyal users were worth.

People wanting cheaper groceries doesn't tell you how you'll cover the cost of helping them.

The people who download a grocery savings app are the least likely people in the country to pay a subscription. And no supermarket is paying you an affiliate fee to send customers to the other supermarket.

I'd leave this idea alone unless I had a specific problem the others haven't solved.

Start with a customer you can reach and work you understand. Being able to build the app is a fairly thin reason to start a business.

Swipe butter

A useful comparison needs the right product, pack size, store, price and collection time.

Then come the club prices, multibuys and substitutions.

A cheaper basket is fairly easy to produce if you quietly leave out the items you couldn't find.

In an ecommerce catalogue, two similar names can describe different products.

Two completely different names can describe the same one.

The customer finds the mistake after you've congratulated yourself on the matching rate of 99% because claude said so.

ShelfScout says it collects prices from each retailer's online catalogue, by store. It also says shelf prices can differ.

That's a useful admission.

An online price only what the retailer listed, but it has 0 context on the clearance sticker next to it.

Online is about 8% of the big supermarkets’ sales. Most of what people actually pay never shows up in any of these databases.

Basket Brain's developer describes running a hidden Firefox browser to get login cookies before collecting prices.

That's one project's approach, others no doubt will still come along and do it ‘differentiated’. But all of these need someone to keep it running.

I wrote about vibe coding making the first build cheap.

This is where the ongoing bill turns up. The shopping list can look finished while the data needs another morning of attention.

Same thing is happening in invoicing tools, rostering apps and AI note takers. Cheap to build the front.

Same expensive plumbing underneath, built twenty times.

I'd rather share that work

The two GitHub projects in the list are a better starting point than another blank repository.

Jason's supermarket-prices project links to KiwiPrice as its live demo.

JavaGT's nz-grocery-prices collects and archives prices, and separates confirmed product matches from fuzzy suggestions.

I'd put more faith in these builders, and even better if it worked together.

Pool the product matching, price collection and maintenance. Give people fewer broken comparisons and a reason to come back.

I'd like to see a maintained product matching table, store IDs and dated price observations that other builders can use. Somebody would have to fund it and check it.

Open source doesn't pay the hosting bill by itself.

The duplication isn’t an accident. In 2022 the Commerce Commission said supermarkets should cooperate with price comparison services.

Should.

Would.

Wont.

Since then Foodstuffs has removed sort by price from the New World and PAK’nSAVE websites.

Better still, get dependable feeds from the retailers.

New Zealand is already moving towards required pricing data access in electricity: MBIE's designation factsheet sets out specified product data due from 1 July 2027.

The law to do it, the Customer and Product Data Act, is already passed. It’s pointed at power plans, which most of us look at once a year.

Not groceries, which we buy every week.

There is still room for different apps.

Meal planning, shopping online and scanning things in the aisle are different jobs. Sharing the data work would give each of them a better chance of being useful.

Twenty people separately checking the same butter seems a poor use of everyone's Saturday.

So What.

If you're building the next one, show me a basket with a missing product, a member-only special and a price that hasn't refreshed since Tuesday. Tell me who fixes it.

I can already download a shopping list. What I want is to trust the total before I leave the house.