基于边缘计算的实时绩效

We’ve gone through 20% of the 21st century. It’s safe to say digitalization isn’t a new concept anymore. Things are fully or at least mostly online, and they tend to escalate in the digital direction. That’s why it’s important to keep up. Opportunities or mistakes could have tremendous effects depending on it.

我们经历了21世纪的20％。可以肯定地说数字化不再是一个新概念。事物完全或至少大部分是在线的，并且它们倾向于在数字方向上升级。这就是为什么保持跟进很重要的原因。机会或错误可能会因此而产生巨大影响。

At the heart of any digital strategy, there’s performance optimization. Because the impact of your strategy, positive or negative, could come at a much larger scale and faster pace in the digital world, you need to always fine-tune its performance. This is especially important in digital marketing, where low performance means burning marketing budgets and get no customer.

性能是任何数字策略的核心。由于您的策略的正面或负面影响可能会在数字世界中以更大的规模和更快的速度出现，因此您需要始终调整其性能。这在数字营销中尤其重要，在数字营销中，低绩效意味着要消耗大量营销预算并且没有客户。

The sad truth about this is: everybody knows this, but not many know how to practice it. They would end up making expensive, consequential mistakes that lead nowhere. To help you fix them before ever making them, below are the three most critical mistakes in a digital marketing strategy that you need to identify — along with my recommendations for efficient ways to combat them.

可悲的事实是：每个人都知道这一点，但很少有人知道如何实践。他们最终将犯下昂贵的后果性错误，导致无济于事。为了帮助您在制造之前修复它们，以下是您需要确定的数字营销策略中的三个最严重的错误-以及我提出的与之抗衡的有效建议。

1.依靠肠胃的感觉 (1. Relying on Gut Feelings)

The number one enemy of all things performance-related is the negligence of data. Decisions are still often made because somebody has been in the field for decades, or because it feels good, or worse, based on no reason at all.

与性能相关的所有事情的头号敌人是数据的疏忽。仍然经常做出决定是因为有人已经在该领域工作了数十年，或者是因为完全没有理由而感到好坏。

To exemplify, for a website’s cookie consent banner, when torn between a short, snappy headline and a detailed, sincere one (see example below), many would decide to use one and discard the other based on how it looks, sounds and feels personally. “It looks too long”, “it sounds boring”, or “it feels robot-like, not human”, for one or the other, are some typical arguments provided for justification.

例如，对于一个网站的Cookie同意标语，当它在简短而活泼的标题和详尽而真诚的标题之间(见下面的示例)被撕裂时，许多人会根据其外观，声音和个人感觉决定使用一个而放弃另一个。 “看起来太长”，“听起来很无聊”或“感觉像机器人，而不是人类”，这是为辩解提供的一些典型论点。

Chances are, they don’t even know how many customers would agree, what opt-in rate can be expected from each variation, or why a shorter or longer headline even makes a difference.

可能的是，他们甚至不知道会有多少客户同意，每个变化可以预期的选择加入率，或者为什么更长或更短的标题甚至会有所作为。

This form of data-disregarding decision-making is particularly costly and dangerous in digital projects. It simply doesn’t rely on any factual indication that it will work effectively. It’s like driving blindfolded, where disasters will just wait to happen. So, don’t let it. Be data-driven.

这种无视数据的决策形式在数字项目中尤其昂贵且危险。它根本不依赖于任何有效的迹象表明它将有效地工作。这就像蒙着眼睛开车一样，那里的灾难将等待发生。所以，不要让它。受数据驱动。

数据可以来自多种来源，并且可以多种形式出现 (Data can come from multiple sources and in several forms)

For the sources, you can either rely on primary or secondary data:

对于源，您可以依赖主要或辅助数据：

Primary data: The kind of data you or your team can collect directly from the customers, competitors, industries, and other stakeholders. It comes from first-hand research, via surveys, interviews, etc., that you have to manage yourself. Hence, it consumes more time and demands more research expertise. On the other hand, you have more control over the design and methodology.
主要数据：您或您的团队可以直接从客户，竞争对手，行业和其他利益相关者那里收集的数据类型。它来自第一手研究，需要通过调查，访谈等来管理自己。因此，它消耗更多的时间并需要更多的研究专业知识。另一方面，您可以更好地控制设计和方法。
Secondary data: Any data from external, third-party sources like university research, company reports, and the Internet. It’s convenient and free, yet most reliable insights are usually hidden in the vast archives of open data sources. Thus, finding a relevant piece of secondary data could be a challenge, let alone an up-to-date one. This type of data is useful if you prioritize efficiency over quality, and understand which sources are unreliable to avoid.
次要数据：来自外部第三方来源的任何数据，例如大学研究，公司报告和Internet。它既方便又免费，但是大多数可靠的见解通常隐藏在大量开放数据源档案中。因此，找到相关的辅助数据可能是一个挑战，更不用说最新的了。如果您优先考虑效率而不是质量，并且了解哪些来源不可靠，则这种类型的数据很有用。

In terms of data forms, we have quantitative and qualitative data.

在数据形式方面，我们有定量和定性的数据。

Quantitative: Numbers, percentage, date and time, and other similar insights that help you measure how much or how many for a topic. The most common example of quantitative data is KPIs (key performance indicators), where effectiveness is quantified into an amount (e.g. 1,000), proportion (15%), rate (+20%), and so on. It gives you simplicity in understanding how you’re doing but takes away the reasons behind it.
定量：数字，百分比，日期和时间以及其他类似的见解，可帮助您衡量一个主题的数量。定量数据最常见的例子是KPI(关键绩效指标)，其有效性被量化为数量(例如1,000)，比例(15％)，比率(+ 20％)等。它使您可以轻松地了解自己的工作方式，但可以消除其背后的原因。
Qualitative: The said reasons, or complex relations between things and people behind the numbers. Quantitative data provides descriptions, and qualitative data gives you explanations for them. For instance, you see a 10% increase in online traffic after a marketing campaign, but it involves a mix of many assets and messages, so you don’t know what exactly drives the uplift. You can’t decide what to keep, improve, or remove as a result. This question can easily be answered if you ask customers how they feel about which aspect of the campaign, and how it influences their actions. No more blindfolds and assumptions — that’s the power of qualitative data.
定性的 ：所说的原因，或数字背后事物与人之间的复杂关系。定量数据提供描述，而定性数据为您提供解释。例如，您在进行营销活动后看到在线流量增加了10％，但这涉及许多资产和消息，因此您不知道是什么真正推动了这种增长。因此，您无法决定要保留，改进或删除的内容。如果您询问客户他们对广告系列的哪个方面以及它对他们的行为有何影响，您可以轻松回答该问题。不再有眼罩和假设-这就是定性数据的力量。

2.低估实验 (2. Underestimating Experiments)

As a rule of thumb, there’s no optimization without experimentation. It applies the same way to digital performance. If you manage a mobile app, you need to experiment (via A/B tests, smoke tests, beta tests, dogfooding, etc.) different features as well as UX and UI elements to estimate which will work best. If you run a Facebook Ad campaign, you need to test and benchmark different images, texts, and videos against each other as well.

根据经验，没有实验就没有优化。它对数字性能采用相同的方法。如果您管理移动应用程序，则需要尝试(通过A / B测试，冒烟测试， Beta测试，狗食测试等)不同的功能以及UX和UI元素，以评估哪种功能最有效。如果您运行Facebook广告活动，则还需要相互测试和比较不同的图像，文本和视频。

What’s more, testing isn’t only about concrete elements like product features or marketing assets. It’s about learning through idea validation. If you have an idea or a hunch, you can elaborate it into a hypothesis and create an experiment to test that hypothesis. Depending on the result, you can accept or reject that hypothesis, which answers if you have a justified, data-backed reason to rely on your idea. If you don’t test, you’ll never learn.

而且，测试不仅针对诸如产品功能或营销资产之类的具体要素。这是关于通过思想验证进行学习。如果您有想法或直觉，可以将其阐述为假设，然后创建实验以检验该假设。根据结果，您可以接受或拒绝该假设，如果您有正当的，有数据支持的理由来依赖您的想法，则该假设会得到回答。如果不考试 ，就永远不会学 。

What saddens me is the fact that many digital projects or product managers still haven’t adopted a testing mindset. Experimentation isn’t prioritized or even thought of often enough. Hence, you miss countless opportunities to put your ideas under pressure before the customers do. You miss the chance to gain valuable insights for continuous improvement from test results as well. If you’re still making this mistake, it’s time to stop.

让我感到难过的是，许多数字项目或产品经理仍然没有采用测试心态。没有优先考虑实验，甚至没有足够频繁地考虑。因此，您会错过无数机会，使您的想法先于客户承受压力。您也会错过从测试结果中获得持续改进的宝贵见解的机会。如果您仍然犯此错误，那么该停止了。

以科学的过程开始测试 (Start testing with a scientific process)

Experimentation is easier than it sounds. All you need to do is follow the strict science behind it so the insights are clear, correct, and legitimate. It ultimately prevents you from learning the wrong things.

实验比听起来容易。您需要做的就是遵循其背后的严格科学知识，以便洞察力清晰，正确和合法。它最终会阻止您学习错误的东西。

There are many publicly available methods and processes you can follow or adapt from to suit your strategy. A few great examples for these include Optimizely’s model for website conversion rate optimization (CRO) and Customlytics’ framework for app store CRO. There are also tools you can use to operate tests properly, many of which are offered by teams of experts who also provide step-by-step guidance as well. All that’s left to do is test one of them (yes, you can test on the meta-level as well) and adjust or scale overtime.

您可以遵循或适应许多适合您的策略的公开可用方法和过程。其中一些很好的例子包括用于网站转化率优化(CRO)的Optimizely 模型和用于应用商店CRO的Customlytics 框架。您还可以使用一些工具来正确运行测试，其中许多工具是由专家团队提供的，他们也提供逐步指导。剩下要做的就是测试其中之一(是的，您也可以在元级别上进行测试)，并调整或扩展加班时间。

超越测试：创意验证 (Beyond testing: idea validation)

Idea validation helps you learn and predict success or failure before it happens based on the signals from a sample or test subject. Despite the crucial role of experiments in idea validation, they aren’t the only component that counts. There are several more groups of methods you can use to validate an idea, as seen in the following summary:

创意验证可帮助您根据来自样本或测试对象的信号，在成功或失败之前学习和预测成功或失败。尽管实验在想法验证中发挥着至关重要的作用，但它们并不是唯一重要的组成部分。您可以使用更多方法组来验证一个想法，如以下摘要所示：

What’s next: pick the best method (or a mix of multiple), test it, and start your learning journey.

下一步：选择最佳方法(或多种方法)，进行测试，然后开始学习之旅。

3.忽略迭代的重要性 (3. Overlooking the Importance of Iteration)

The nature of digital projects is that they must involve the creation of digital assets of some sort in order to operate. In display marketing, it could be an ad banner, infographic, or YouTube video. In Search Engine Optimization (SEO), it includes the keywords used in content writing and search ads bidding. In mobile app development, UI elements like icons and status bars are created. In any case, it’s vital to produce, develop, and manage your digital assets.

数字项目的本质是，它们必须涉及某种形式的数字资产的创建才能运行。在展示广告营销中，它可以是广告横幅，信息图或YouTube视频。在搜索引擎优化(SEO)中，它包括内容撰写和搜索广告出价中使用的关键字。在移动应用程序开发中，将创建UI元素(例如图标和状态栏)。无论如何，生产，开发和管理您的数字资产至关重要。

The most common and dangerous behavior that many still follow when it comes to digital assets is being too defensive about each of their versions. On the one hand, it’s understandable that all assets must be treated with respect. After all, they’re the resources that fuel your strategy and the materials that help it make an impact. On the other hand, failing to let go could risk making you closed-minded and inflexible.

在涉及数字资产时，许多人仍然会遵循的最常见和最危险的行为是对每个版本的防御性过强。一方面，必须尊重所有资产，这是可以理解的。毕竟，它们是为您的策略提供动力的资源，以及可以帮助您产生影响的材料。另一方面，不放手会冒使您头脑笨拙和僵化的风险。

Since nobody is perfect, whatever kind of strategy you come up with will most likely always have space for improvement. You need to keep iterating, and it won’t work unless you’re flexible enough, as you’ll grow resistant to change. This is a dreadful mistake that you need to avoid as early as possible — and remembering the following will help:

由于没有人是完美的，因此无论您想出哪种策略，都很可能总有改进的空间。您需要保持迭代，除非您有足够的灵活性，否则它将无法正常工作，因为您将变得难以抗拒变化。这是一个可怕的错误，您需要尽早避免-记住以下内容将有所帮助：

除非所有资产都是100％优化的，否则它们都是原型 (All assets are prototypes unless they are 100% optimized)

To make sure you’re always up for iterating, remember the true nature of all digital assets: they’re temporary. Like prototypes, as soon as an updated, fine-tuned, and superior version is available, they’ll be “killed.”

为确保您始终愿意进行迭代，请记住所有数字资产的真实性质：它们是临时的。与原型一样，一旦有更新，经过微调和更高级的版本可用，它们就会被“ 杀死” 。

By being defensive and resistant to change, you won’t allow yourself to learn and your assets to “evolve.” By contrast, if you let go of them and iterate rapidly, you’ll give the chance to quickly become the best they can be.

通过具有防御能力和抵御变化的能力，您将无法学习，资产也无法“发展”。相比之下，如果您放开它们并快速迭代，您将有机会Swift成为最好的它们。

单一资产的影响不如许多资产的混合 (Single assets are not as impactful as mixtures of many)

Having the best assets of all types isn’t enough. You also need to be sure they work well collectively. Take app store CRO campaigns, for instance, where you have a range of on-page text and visual assets (e.g. title, descriptions for text and icon, screenshots for visual) at your disposal. Even if you have the best ever version of everything, chances are they won’t fit well together straight away. Their synergy or relationship will likely show space for improvement.

拥有所有类型的最佳资产还不够。您还需要确保它们共同运作良好。例如，以应用商店CRO活动为例，您可以随意使用一系列页面上的文字和视觉资产(例如标题，文字和图标说明，视觉截图)。即使您拥有所有版本中最好的版本，它们也可能无法立即很好地融合在一起。他们的协同作用或关系可能会显示出改进的空间。

If you ignore them, and they tell contradicting stories that confuse the audience, it becomes a marketing nightmare. Entire campaigns could be ruined. In cases like this, having wonderful individual assets may still lead to uninspiring results. Make sure all digital assets you work with get along well like a team.

如果您忽略它们，并且它们讲述的矛盾故事使观众感到困惑，那么它将成为营销的噩梦。整个竞选活动可能会被摧毁。在这种情况下，拥有出色的个人资产可能仍会导致令人鼓舞的结果。确保与您合作的所有数字资产都像团队一样相处融洽。

Digitalization involves a high dose of uncertainty and unpredictability. In the online environment, you can’t walk up to a customer and talk to them, or dissect a software to see how each component works, or hear the claps coming from readers who enjoy your blog posts. It’s a grey zone full of traps, where you make mistakes that cost you a piece of business at a time.

数字化涉及大量不确定性和不可预测性。在在线环境中，您无法走近客户并与他们交谈，也无法剖析软件来查看每个组件的工作原理，也无法听到喜欢您博客文章的读者的鼓掌。这是一个充满陷阱的灰色地带，您在这里犯错，一次就要付出一笔生意。

Among those mistakes, there are three that I believe to have greater significance than the rest:

在这些错误中，我认为有三个比其他错误具有更大的意义：