Which elements you can personalize with vidflare

The essentials at a glance

• Almost every element in a video can be controlled by data: text, font, color, images, video clips, audio, text-to-speech, map sections, position, and size.
• Feeds can also be used to determine whether an element appears at all and which source or text is used.
• For sophisticated motion design, masks, alpha and luma mattes, keyframes, interpolations, as well as wiggle and oscillator effects are available.
• Personalization is most effective in CRM, where you already know your customers and their specific situation.
• Personal data in video is a data protection matter. Clarify the use with your data protection officer in advance.

How personalization works technically: template plus dataset

It starts with a master template. You build the video once in the web-based editor and define which parts are variable. You link these parts to data points from a source, such as an API, a spreadsheet, or a JSON file. A unique video is then generated for each dataset (vidflare.io).

The practical effect: If something changes, you don't have to touch the video again. You update the data, and the affected versions are recalculated. You receive a feed that corresponds to the one you imported, now supplemented with the video URLs. You can integrate these URLs directly into your email sequences, customer platform, or campaign.

The variable elements are rendered directly into the video rather than being placed as an overlay. For the viewers, a personalized version looks exactly like a professionally produced film.

The more interesting question is what can be variable. Here are the individual layers.

Text and typography: the easiest entry point

Text is the layer where most projects start, and for good reason. You don't need new footage, just data that is already in your CRM. Names, contract types, product names, remaining terms, bonus points, appointments, or personal voucher codes can be inserted directly into titles, lower thirds, or end cards.

It gets exciting when not just the content varies. Font, font size, and text color can also be pulled from the data. You don't have to worry about long surnames or product names with many characters: vidflare automatically adjusts the text size to the available space so that nothing breaks awkwardly or gets cut off.

This is also interesting for companies with multiple brands or dealer networks. An insurance company with regional agencies can use the same template and insert the address of the relevant agency and the name of the personal contact for each customer, without having to build a separate video for every agency.

Swap images, video clips, and colors

On a visual level, you can switch image and video sources via data points. For an online shop, this means: the video displays the product a user recently viewed or purchased, using the image from the product catalog. For a car dealership: the purchased model, ideally in the correct trim. Individual scenes can be swapped just as easily, such as using a different intro clip for new customers versus long-term ones.

The effort involved depends heavily on how much material you already have. Many companies possess more usable images than they realize: product photos, branch photos, and corporate videos that can be edited into short clips. Start with this existing inventory. Creating custom footage for every variation is only worth it once you know which versions actually drive results.

Color schemes are also controllable, and this applies to the entire motion design. It sounds like a minor detail, but it is often the fastest way to adapt a video to a specific sender: dealer colors, tariff-specific colors, or a seasonal variation. On top of that, you can adjust the dimensions and positions of elements. A logo can appear larger or smaller depending on the dataset, or an image can shift to a different spot if more space is needed for text next to it.

Sound and speech: Text-to-Speech and audio sources

A name on the screen catches the eye. A name that is spoken aloud feels significantly more personal. With text-to-speech, text from a dataset becomes an audio track. The voiceover can greet the customer by name, mention their tariff, or announce an appointment without anyone having to step into a studio for every single variation.

In practice, it is worth spot-checking the audio output before sending. Rare names, abbreviations, or product names containing numbers are not always pronounced as you might expect. A few test datasets with tricky cases will save you trouble later on.

In addition to speech, you can also select audio sources from your data. You might consider different music based on the target audience or a pre-recorded sentence that fits a specific segment. If the voice quality is particularly important to you, you can combine both: professionally recorded base tracks and individual variable sections via text-to-speech.

Map snippets: when location matters

Map snippets can also be generated from data. This is useful whenever a specific location is involved: the nearest branch, the responsible workshop, a pickup point for an order, or the location of a consultation appointment.

An energy provider could show new customers where their customer center is located in a welcome video. A bank could show all affected customers the new location relative to the old one during a branch relocation. This is much more intuitive than an address buried in the body text of an email.

You should be especially careful with customer address data. Showing a branch is unproblematic. Showing a customer's home address on a map in a video is not, especially if the video is accessible via a link that can be forwarded.

Feeds control what happens in the video

So far, we have discussed content that gets swapped out. At least as important is the question of whether an element appears at all. Through connected feeds, you can provide every element with four types of instructions:

• On/off: An element or an entire scene is only shown if the dataset dictates it. A mention of the loyalty program only appears for customers who are not yet members.
• Source: The element is assigned a different source, such as a different image, clip, or audio file.
• Text: The content of a text element is set from the feed.
• Properties: Values such as color, size, or position are pulled from the feed, as described above.

The on/off control essentially turns a template into a modular kit. Instead of creating ten videos for ten segments, you build one template with all the scenes and let the feed decide which ones a specific customer sees. For example, a car insurance provider can show a scene about no-claims bonuses only to customers with a long accident-free history, and a scene about second-car tariffs only to households with one vehicle.

Feeds don't have to come from a CRM. A product feed from your shop works just as well, for instance, to keep prices or availability up to date. If the feed changes, the affected videos are recalculated.

Tools for motion designers: mattes, masks, keyframes

Personalized videos often have a reputation for looking like they were built from a kit. This is usually because variable elements are rigidly placed into a fixed layout. For dynamic content to keep up in terms of design, you need tools that motion designers are familiar with from their usual software. The vidflare editor includes the following, among others:

• Masks: Content is cropped to a shape, such as a product image inside a circle or a clip behind a text area.
• Alpha and luma mattes: An element uses its transparency or brightness to determine which parts of another element are visible. This allows a variable image to be blended directly into a piece of text.
• Keyframe animations and interpolations: Movements, scaling, and transitions are defined via keyframes. Interpolation determines whether a movement runs linearly or accelerates and decelerates smoothly.
• Wiggle and oscillator: These create random jitter or steady oscillations, for example, for a pulsing call-to-action element or a slightly lively camera movement.

For you as a marketing or CRM manager, this means one thing above all: your template doesn't have to look any simpler than your other videos. If you have previously worked with an agency or an in-house motion team, you can involve them directly. They build the template using their usual tools, and you provide the data.

Along the customer journey: where different elements fit

In CRM, you have a major advantage over advertising to strangers: you know who the person is, what they have bought, and where they currently stand. That is why it is worth assigning every element to a specific moment in the customer relationship. Here are a few examples from industries typical of our readers:

Welcome and onboarding. After taking out car insurance, the customer receives a video that greets them by name (text and text-to-speech), mentions their vehicle and tariff, and explains only the benefits they have actually purchased (on/off control). At the end, a map shows the nearest partner workshop. The insurance policy remains, but the video makes it easier to understand.

After the purchase. One week after delivery, the online shop sends a short video with care tips for the purchased product (using images and clips from the product feed) and displays matching accessories. Customers who have already purchased accessories will not see this scene.

Service occasions. The car dealership sends a reminder for an inspection, including the model, mileage from the last service visit, appointment details, and a map. A version with different text and color schemes can be used for customers whose warranty is about to expire.

Renewal and upgrade. Shortly before a contract expires, an energy provider displays the customer's previous consumption as animated bars—with lengths derived from the dataset—and introduces a suitable follow-up tariff. The values come from the dataset, while the animation is generated from the template.

Loyalty and special moments. Annual reviews, customer anniversaries, or bonus program status updates are occasions where customers perceive a personalized video as a thoughtful gesture rather than advertising. In these cases, the design and tone can be a bit more emotional.

Retention. For customers who haven't been active for a while, the approach can be tiered based on data: different intros and different offers, depending on how long ago their last order was.

The occasion is the best guide for determining which elements to use. A video that simultaneously varies name, location, product, color, and voice can quickly feel cluttered. Our experience shows that two or three well chosen personal details are more effective than a fireworks display of data.

The same elements in paid media

In advertising campaigns, you usually don't know the individual person. Personalization here is based more on segment, region, product, or context. The elements remain the same; only the data source changes.

A product feed provides the image, price, and product name for every ad. Regional campaigns display the appropriate branch on a map and the local contact person in the text for each delivery area. For digital out-of-home, the feed can update daily offers without anyone needing to export new spots.

Then there is the question of format. Campaigns often require the same creative in many aspect ratios and resolutions for social media, connected TV, in-banner, or in-store screens. Because dimensions and positions are adjustable, a template can be designed to cover these variations. And because variations require little effort, you can A/B test more of them against each other.

Consider data protection from the start

As soon as names, contracts, consumption figures, or addresses are included in a video, you are processing personal data under the GDPR. You need a legal basis for this, and the usage should align with what your customers expect and what they have consented to.

You should clarify a few questions before starting: Which data fields does the video really need? Where is the data processed and where are the finished videos stored? Who can see a video if the link is forwarded? And how long will the videos remain accessible? Regarding the first question, the principle of data minimization applies: what is not necessary for the message does not belong in the dataset.

Regarding processing: According to their own statements, vidflare hosts and operates all system components in its own data center in Hamburg (vidflare.io). This does not replace a case-by-case assessment. This article does not constitute legal advice. Involve your data protection officer early on, ideally as soon as you start selecting data fields.

Sources

• vidflare: Platform for scalable video personalization, accessed on October 6, 2026 (functional scope, data sources, hosting; manufacturer information)
• Information on text-to-speech, map snippets, feed control (on/off/source/text) as well as masks, mattes, keyframes, interpolations, wiggle, and oscillator: internal vidflare product information