If you sell electronics, almost all of your profitable Google Shopping traffic comes from one kind of query: the exact model number. Someone typing WF-C710N price or Lenovo IdeaPad Slim 3 16GB has already done the comparison and is looking for a seller. Someone typing best laptop under 60000 is three weeks away from buying anything. Google Shopping for electronics works when your feed is built to win the first query and your campaign structure is built to avoid paying full price for the second.
That's the whole thesis. The rest is execution, and in electronics the execution is unusually fiddly because of identifiers, variants, brand authorisation and margins that don't forgive a sloppy bid.
Your feed is the keyword list
There is no keyword field in Shopping. The title, description, brand, GTIN and product type are what Google matches against a query, and the title does most of the work. For electronics the pattern that performs is brand, then model name, then the model code, then the spec that distinguishes the variant, then colour.
So instead of Premium Noise Cancelling Wireless Headphones, you want something closer to Sony WH-CH720N Wireless Noise Cancelling Headphones, Black. The model code matters more than anything else in that string, because the model code is what people type. On laptops and phones, put the configuration in the title: 8GB/512GB, 256GB, 55-inch, 4K. Those are query tokens. Buyers use them.
Roughly the first 70 characters show in most placements, so front-load. You get up to 150, and the rest still counts for matching even when it's truncated on screen.
One thing we fix on nearly every electronics feed we inherit: variant titles that are identical across storage or colour. If three variants of the same phone share a title, Google can't tell them apart, your reporting is useless, and you end up bidding against yourself on the version with the worst margin.
GTIN, MPN and the disapprovals you'll actually hit
Electronics are the one category where Google assumes a valid GTIN exists, because it usually does. Submit a wrong one and you don't get a soft warning, you get the product pulled. Three failure modes we see repeatedly:
- The GTIN belongs to the base model. Someone copies the 128GB barcode onto the 256GB variant. Google matches it to the wrong product, your price looks absurd against the benchmark, and the offer gets flagged.
- Brand restriction. Several manufacturers restrict who can advertise their products. If you're an unauthorised reseller or you're selling parallel imports, expect disapproval under counterfeit or policy grounds, and expect the appeal to go nowhere. This is a sourcing problem, not a feed problem.
- Price mismatch between feed and page. Your theme shows a different price after a cart discount or a currency rule, Google's crawler sees it, and the SKU drops out. Keep the structured data on the product page honest and matched to what the feed sends.
If you sell refurbished or open-box, set the condition attribute properly and say so in the title. Trying to pass refurbished stock as new is the fastest way to lose a Merchant Center account, and the refurbished buyer is a genuinely different, often less price-sensitive customer once you've told them what warranty they get.
The arithmetic that decides whether you should bid at all
Electronics margins are thin enough that you can run a campaign at a 6x ROAS and still lose money. Work it per order before you set a budget.
Take a ₹12,999 pair of ANC headphones at 16% gross margin, so ₹2,080 of gross profit. On model-number searches you might pay ₹22 a click and convert the product page at 1.8%. Customer acquisition cost is 22 ÷ 0.018 = ₹1,222. That leaves ₹858. Now subtract ₹90 shipping and roughly 2% payment gateway on ₹12,999, which is ₹260. You're at ₹508. Allow a 6% return rate with ₹150 of reverse logistics each time and you land near ₹470 of contribution per order, about 3.6% of order value.
Thin, but positive, and it scales.
Now run the same product against the generic query. CPC drops to ₹14 because the auction is less contested, but page conversion falls to 0.4%. CAC = 14 ÷ 0.004 = ₹3,500 against ₹2,080 of margin. You lose ₹1,420 on every order and the blended ROAS hides it, because the model-number conversions are subsidising the category-term ones in the same campaign.
This is why generic-term Shopping traffic in electronics is overrated. It's not useless, it's just a top-of-funnel spend that should be budgeted and measured as one, not mixed into a performance campaign and called efficient.
Structure campaigns around margin, not around product type
Custom labels exist for exactly this. We typically split a catalogue into four bands: high-margin accessories, mid-margin own-brand or exclusive SKUs, commodity SKUs where you're within a few hundred rupees of the market, and SKUs where you're simply not price-competitive.
That last band is the interesting one. If an authorised marketplace seller is ₹800 under you on a ₹30,000 SKU, you will not win that auction profitably, no matter how good the creative is. Either exclude it from paid and let it sit in free listings, or change the offer: bundle a case and a two-year warranty extension so the comparison isn't like-for-like.
On campaign type, the honest trade-off: Performance Max gets more volume and handles the long tail of model codes better than anything you'd build manually, but it gives you search categories rather than a proper search terms report. Standard Shopping gives you the search term data and real negative keyword control. For electronics catalogues with a few hundred SKUs we often run both, with Standard Shopping carrying the top 30 revenue SKUs so we can see and block the queries that waste money, and PMax catching everything else. PMax takes priority when both could serve, so the standard campaign effectively becomes a diagnostic layer you read rather than a volume driver.
If you also sell into the UK or the EU, a Comparison Shopping Service partner lowers your effective cost per click on those storefronts. It does not apply in India. If an agency offers you a CSS saving on Indian traffic, that's a reason to end the conversation.
What has to be true on the product page
High-intent Shopping clicks land on a product page with a buyer who already knows the spec. They are checking three things: is it genuine, when does it arrive, and what does it cost per month.
- EMI, visible above the fold. Above roughly ₹15,000 ASP in India, no-cost EMI messaging changes the decision. A Razorpay affordability widget that reads "from ₹1,249/month" next to the price does more for conversion on a ₹14,999 SKU than any amount of trust-badge decoration.
- Warranty terms in plain words. Brand warranty or seller warranty, how long, and who services it. Electronics buyers read this.
- GST invoice. A surprising share of mid-ticket electronics demand is small businesses buying a monitor or a printer. Saying you issue a GST invoice with the buyer's GSTIN is a genuine differentiator and costs nothing.
- Delivery date, not a shipping policy link. "Delivered by Thursday to 560068" beats "ships in 2-5 days".
On COD: for anything above roughly ₹10,000, COD on cold Shopping traffic is a return-to-origin machine. We'd rather limit COD by cart value and offer a 2-3% prepaid discount, which costs less than the RTO it prevents. If you must keep COD open, put an OTP confirmation step in and watch the RTO rate by campaign.
Page speed and spec filters do more than bid tuning
Electronics product pages are heavy by nature: spec tables, comparison charts, six image angles, a video, a review widget. We've seen LCP sit above 4 seconds on category pages purely from unoptimised hero images and three review apps loading in sequence. Every tenth of a second there is money, because you're paying ₹22 to land the click. Fixing the theme is usually cheaper than fixing the bids, and the gain is permanent. That's the bulk of what our Shopify speed work does on electronics stores.
The second post-click lever is navigation. Shopping traffic doesn't always land on the exact model, and a buyer who arrives on a 65-inch TV page but wanted 55-inch needs to get there in one click, filtered by panel type, resolution and price band. Shopify's native filters handle simple cases; once you're filtering across RAM, storage, screen size, connectivity and brand at the same time, you need something purpose-built. FilterPro is the app we built for that, and spec-level filtering is where it earns its keep on an electronics catalogue.
You're being price-benchmarked whether you look or not
Merchant Center shows price competitiveness and best seller data for your categories. Read it monthly. Two things come out of it: SKUs where you're above the benchmark and quietly burning budget, and SKUs where you're well below it and could take ₹300 of price back without losing the sale.
Use the sale price attribute properly rather than editing the base price. Google can show a price-drop annotation on the listing when it has a price history to compare against, and that annotation does real work on commodity SKUs. Overwriting the compare-at price every week trains the system that your discounted price is your normal price.
Festive peaks change the maths, not the strategy
From late September through Diwali, electronics CPCs in India climb while marketplace discounting compresses your price position at the same time. Plan for both. The approach that holds up: pull budget forward into the two or three weeks before the marketplace sale events, when intent is rising and auction pressure hasn't peaked, then shift to accessories and own-brand SKUs during the sale window itself when the commodity categories are unwinnable.
Set your feed up before the rush. Inventory sync lag of even a few hours during a peak means you pay for clicks on sold-out SKUs, and repeated availability mismatches get your account flagged at the worst possible time.
Where to start this week
Export your feed. Sort by clicks over the last 90 days and look at the top 50 SKUs only. Check that each one has a correct, variant-specific GTIN, a title that leads with brand and model code, and a page price that matches what the feed sends. That single pass usually moves more revenue than a month of bid adjustments, and you can do it yourself in an afternoon.
If you'd rather have someone else look at the feed, the account structure and the product pages together, our team does a free store and feed audit, and if the fixes turn out to be theme-level you can bring in a developer for the build without committing to a retainer.