Types of online advertising — ways to promote a company

Internet advertising works best when you match the format to the user’s intent and measure the effect in specific events, not “by eye”. Search engine campaigns (SEM) let you capture demand exactly at the moment someone enters a query with a purchase or contact need. In turn, social media advertising works brilliantly when you want to build demand and reach people who are not yet actively looking for a product. In both channels, the result depends on: correct measurement setup (e.g. GA4, Pixel/CAPI), the quality of the creatives, and the alignment of the landing page with intent. In this section, you get practical tips on how to set up and optimise Search campaigns and how to run effective campaign optimisation in social media. Keep reading if you want to reduce irrelevant clicks and get to conversions faster.

How to use Search campaigns in SEM effectively?

You will use Search campaigns in SEM most effectively when you set ads against specific keywords that respond to a “here and now” need and when you attribute results by clicks (CPC) and conversions in GA4. This format works because you reach a user who is already entering a query such as “dentist Kraków” or “online OC insurance” and expects a quick answer.

To increase impression share and reduce traffic “poaching”, use brand ads when competitors bid on your company name. In reports, monitor among other things Search Impression Share and the cost of lost clicks, so you can see where results are really leaking away.

Most of the budget in Search is usually “lost” not because of bids, but because of query mismatches and a lack of exclusions. If you want to cover the long tail of phrases without manually selecting keywords, use DSA (Dynamic Search Ads), but make sure you exclude sections such as terms and conditions or the blog, and add negative keywords so you do not pay for irrelevant queries. With match types (broad/phrase/exact), you can maintain quality control through the search terms report and automated rules, e.g. pause at CPA > X. Broad can work well when you have goals, conversion values and an exclusion list set up correctly, because then the algorithm optimises for the real outcome.

• Ad extensions (sitelinks, calls, prices) – add information without increasing the budget and can direct users, for example, to “price list” or “implementation”.
• Conversion tracking – clicks on the phone number from the call extension can be measured as conversions in Google Ads.
• Bidding strategies (tCPA, tROAS, Max Conversions) – they usually make sense with a stable volume (e.g. several dozen conversions per month) and correctly assigned values.

You will use video advertising on YouTube most effectively when you match the format to the objective (sales vs reach) and base assessment on video metrics and conversions, not just clicks. TrueView In-Stream (skippable) works well for scaling and paying for a view/action, while 6s bumpers are ideal for frequent brand reminders with controlled frequency. If you create vertical creatives for mobile, Shorts Ads require a strong start in the first 1–2 seconds and a clear presentation of the product. It is worth combining Shorts performance with view rate, growth in branded searches and assisted conversions.

The most common reason ads are skipped is too slow a start, which is why the product and the benefit should appear in the first few seconds, and the CTA must be unambiguous. In practice, shorter versions help (e.g. 6–15s) as well as clear calls to action such as “book a free quote” or “check availability today”. To measure business impact, combine VTR and CPV with Brand Lift (YouTube) and assisted conversions in GA4. At larger scale, incremental tests (geo or holdout) also make sense, as they show whether the ad generates additional impact, rather than merely “attributing” sales.

YouTube ad sequences work when you want to tell a story and guide the audience step by step instead of bombarding them with a single creative. You can set the order as follows: 6s bumper (reach) → 15s product demo (consideration) → 6s with an offer (conversion), while also keeping an eye on frequency and excluding people who have already bought. This setup lets you control what the user sees first and how quickly they move to the sales message. Assess the effect not only by impressions, but also by later conversions and changes in interest in the brand.

Programmatic and native advertising: how to control delivery and targeting?

You control delivery and targeting in programmatic and native advertising most effectively through a deliberate choice of purchases (marketplace/private deals), brand safety and frequency limits, rather than leaving the campaign “wide open”. In tools such as Display & Video 360 or Adform, you can buy inventory from premium publishers and set brand safety rules, including third-party filters (e.g. IAS, DoubleVerify).

In Google Display Network you can reduce accidental impressions through exclusions of mobile apps, content categories and frequency caps. This means the budget is less likely to “leak” into placements that do not deliver quality.

Following cookie restrictions, contextual targeting is playing an increasingly important role, choosing delivery placements based on the topic of the content rather than tracking the user. Compare contextual with behavioural targeting in tests (e.g. holdout) or through analysis of growth in branded searches. In native advertising (Taboola, Outbrain), advisory content and rankings work best, but the key is to monitor the quality of traffic sources and reject placements with a high bounce rate. If you want to stand out from standard banners, rich media lets you add interactions (e.g. product scrolling or a mini-configurator) and measure not only clicks, but also in-creative behaviour and view-through conversions.

• Brand safety: use domain exclusion lists, content categories and negative keywords, and in awareness campaigns consider IAS/DoubleVerify filters.
• Frequency control: set a frequency cap (e.g. 2–5 impressions per user per day) and observe the impact on CTR, CPA and unique reach.
• Traffic quality: in native placements, reject sources with high bounce rates, and in in-app placements exclude low-quality apps and filter placements to reduce accidental clicks.
• Pricing model aligned to the objective: for awareness, choose CPM/vCPM and measure reach and viewability, and for sales optimise for CPA/ROAS with properly set attribution.

Choose display and programmatic billing models to match the goal, because this channel should rarely be judged solely on clicks. For brand awareness, CPM/vCPM and measurement of reach and viewability make more sense, while for sales-led activity CPA/ROAS with consistent attribution more often matters. In in-app campaigns, you can additionally reduce irrelevant interactions by optimising for viewable impressions (vCPM) or conversions instead of cheap CPC. Such control over delivery and targeting helps maintain quality, even when you scale reach beyond search.

Remarketing campaigns: segmentation and dynamic creatives

Remarketing campaigns work best when you divide audiences into segments by stage in the journey and match each with a separate creative and message. In practice, the main groups are: people who visited the site, viewed a product, abandoned basket, and customers, and each of them needs a different “close” (e.g. delivery or warranty information, rather than a general banner). The biggest difference is that remarketing is not one campaign, but a set of messages matched to intent and the decision stage. This approach reduces wasted budget on repeating the same message to people who are at different stages.

Choose tools according to the sales model and integration needs: Klaviyo is a common choice in e-commerce (segments and RFM), GetResponse provides solid Polish support and landing pages, and HubSpot combines marketing with the sales process also in B2B. To make segmentation better reflect the real value of the customer, combine e-commerce data (e.g. Shopify/WooCommerce) with CRM data (e.g. Pipedrive, HubSpot) and supplement it with offline purchases.

At the same time, keep deliverability under control by configuring SPF/DKIM/DMARC and maintaining list hygiene (e.g. removing contacts with no opens for 90–180 days), because complaints and bounces lower domain reputation. Treat SMS marketing as an “instant” channel for alerts (delivery, 24-hour promotion) with clear consent and a short CTA, measuring revenue per send and unsubscribe rate.

Working with influencers: how to choose and measure effectiveness?

Working with influencers is easiest to optimise when the priority is niche fit and credibility, not just follower count. Micro-influencers (e.g. 10–50 thousand followers) can deliver a better acquisition cost if their community genuinely matches the product and topic. The key criterion is the creator’s alignment with your category and the product use context, because this directly translates into traffic quality and sales. For quick searching and initial profile vetting, platforms such as indaHash, Influtool, Upfluence or local databases (e.g. Whitepress in its influencer section) will be useful, and you can base mention monitoring on Brand24 or SentiOne tools.

It is worth setting the billing model so as to limit the risk of overpaying without certainty of results. In practice, a hybrid variant works well: a lower upfront fee plus a commission on sales, often based on an affiliate link (e.g. Awin, TradeDoubler) or a discount code assigned to the creator.

The brief should cover the main benefit, prohibited wording, required product shots and legal conditions, while still leaving the creator space for their own language, because an overly “rigid” instruction undermines authenticity. It is also worth specifying in the agreement provisions on licensing and whitelisting (e.g. ads from the influencer’s account in Meta), if you plan to use the materials legally in paid campaigns.

You assess collaboration effectiveness reliably when you collect hard sales data and compare it with the baseline period, rather than basing conclusions only on reach and likes. Use discount codes, UTM links and dedicated landing pages, and compare the results in a 7–14 day window after publication, taking seasonality into account. In Poland, materials must be clearly labelled (e.g. “ad”, “advertising collaboration”), and it is good to include the labelling rules in the brief, because responsibility covers both the brand and the creator. Even if the creator does not generate a lot of sales, you can use their content as UGC for Meta/TikTok ads, testing different hooks and captions, and for brand impact consider longer series (e.g. 3–6 publications per month for 3 months).

Marketplaces and price comparison sites: product advertising strategies

You build a product advertising strategy in marketplaces and price comparison sites most effectively when you start with the quality of product data and only then increase budget and reach. In Google Shopping, Merchant Center and the product feed are key: titles with parameters, GTIN and correct categories, because data quality affects impressions and acquisition cost in Shopping and PMax. This format responds to the need to show the price and product without having to build recognition from scratch. If the feed is inconsistent or incomplete, campaigns lose effectiveness faster, because the system has poorer foundations for matching queries and serving ads.

The choice of channel (own store vs marketplace) should result from the goal and acceptance of the “platform cost”. Allegro Ads delivers faster sales at the expense of margin and dependence on the platform, so when promoting specific auctions, monitor profitability (ROAS after margin) and stock levels, because lack of availability burns budget. On Amazon Ads (Sponsored Products/Brands), ads are based on buying intent, but they require refined listings (titles, A+ Content, reviews) and ongoing ACOS/ROAS control depending on margin. In price comparison sites (Ceneo, Skąpiec), traffic is strongly “price-led”, so keep an eye on pricing policy, cost per click and availability, because losing on price or a faulty feed quickly reduces profitability.

You are most likely to maintain the profitability of product ads when you manage the feed and segment activity according to margin and logistics. Feed management tools (Baselinker, Channable, Feedonomics) help map attributes and apply rules, for example automatically adding colours/sizes to titles or excluding low-margin products from campaigns. If you see that ROAS is high but profit is low, split campaigns by margin (high/medium/low) and set different tROAS targets so that you do not promote “loss leaders” that damage financial performance. Additionally, work on reviews and customer service (on marketplaces, quality metrics matter, and in comparison sites the store ratings and delivery/returns conditions are visible) and do not advertise products without stock for 3–7 days of sales, because the promise of availability and SLA directly affects conversion.