營(yíng)銷(xiāo)大數(shù)據(jù)分析 關(guān)鍵技術(shù)

Perhaps more than any other field, marketing, especially digital marketing, revolves almost entirely around data. This makes it a rich and rewarding business to support as an analyst or a data scientist, as the volume and utility of data can be incredibly high, increasing the need and the scope of potential projects for the analytics professional.

營(yíng)銷(xiāo)，尤其是數(shù)字營(yíng)銷(xiāo)，可能比其他任何領(lǐng)域都幾乎完全圍繞數(shù)據(jù)展開(kāi)。 這使得成為分析師或數(shù)據(jù)科學(xué)家成為一項(xiàng)富裕而有意義的業(yè)務(wù)，因?yàn)閿?shù)據(jù)量和實(shí)用性可能非常高，從而增加了分析專(zhuān)業(yè)人員的潛在項(xiàng)目需求和范圍。

Marketing data has some important pitfalls, however, that can de-rail Marketing analytics programs:

營(yíng)銷(xiāo)數(shù)據(jù)有一些重要的陷阱，但是，它們可能會(huì)使?fàn)I銷(xiāo)分析程序脫軌：

• Marketing data is siloed across numerous sources from ad platform, social, and click-stream data to CRM and CDP data.
  營(yíng)銷(xiāo)數(shù)據(jù)從廣告平臺(tái)，社交和點(diǎn)擊流數(shù)據(jù)到CRM和CDP數(shù)據(jù)的眾多來(lái)源中都是孤立的。
• Integrating click-stream and marketing data is challenging without a cohesive tagging and tracking strategy that's understood by both the analytics team and the marketing team. Integrating sales and marketing data, especially in B2B, is rarely done correctly.
  沒(méi)有分析團(tuán)隊(duì)和市場(chǎng)營(yíng)銷(xiāo)團(tuán)隊(duì)都能夠理解的統(tǒng)一的標(biāo)記和跟蹤策略，將點(diǎn)擊流和市場(chǎng)營(yíng)銷(xiāo)數(shù)據(jù)集成在一起就具有挑戰(zhàn)性。 很難正確地集成銷(xiāo)售和營(yíng)銷(xiāo)數(shù)據(jù)，尤其是在B2B中。
• Marketing teams have nearly a dozen different ways to measure marketing performance, meaning creating dashboards and analysis is challenging.
  營(yíng)銷(xiāo)團(tuán)隊(duì)有將近十二種不同的方式來(lái)衡量營(yíng)銷(xiāo)績(jī)效，這意味著創(chuàng)建儀表盤(pán)和進(jìn)行分析具有挑戰(zhàn)性。

In this article, I’ll explain the 3 critical skills that will enable you to overcome these pitfalls.

在本文中，我將解釋使您克服這些陷阱的3個(gè)關(guān)鍵技能。

1.使用API (1. Working with APIs)

SQL and data manipulation skills aren’t going to be enough to effectively get all the data you’ll need to measure marketing programs — unless you want to manually download excel spreadsheets every day. You’ll need to learn how to code with REST APIs to automatically pull ad platform/click-stream/and other marketing data.

SQL和數(shù)據(jù)操作技能不足以有效地獲取衡量營(yíng)銷(xiāo)計(jì)劃所需的所有數(shù)據(jù)-除非您想每天手動(dòng)下載excel電子表格。 您將需要學(xué)習(xí)如何使用REST API進(jìn)行編碼，以自動(dòng)提取廣告平臺(tái)/點(diǎn)擊流/和其他營(yíng)銷(xiāo)數(shù)據(jù)。

The biggest lie that you may have been told is that you need to be a developer to work with APIs. The truth: you don’t. Here are a couple of tips and tricks that will get you 80% of the way there:

您可能被告知的最大謊言是，您需要成為一名開(kāi)發(fā)人員才能使用API??。 事實(shí)是：您沒(méi)有。 這里有一些提示和技巧，可以幫助您達(dá)到80％的目標(biāo)：

• Read my article on How to Pull Data from An API Using Python Requests. This article reviews how to use the Python requests package to pull email data from the Microsoft Graph API. The methods in that article will extend to working with Ad Platform APIs like the Facebook Graph API or Linkedin Ads API.

  閱讀有關(guān)如何使用Python請(qǐng)求從API提取數(shù)據(jù)的文章。 本文介紹了如何使用Python請(qǐng)求包從Microsoft Graph API中提取電子郵件數(shù)據(jù)。 該文章中的方法將擴(kuò)展為使用Facebook Graph API或Linkedin Ads API等廣告平臺(tái)API。

• Learn the PYODBC package. PYODBC allows you to write SQL statements and work with your database using Python. I use PYODBC to insert the data pulled from various APIs into a database.
  了解PYODBC軟件包。 PYODBC允許您編寫(xiě)SQL語(yǔ)句并使用Python使用數(shù)據(jù)庫(kù)。 我使用PYODBC將從各種API提取的數(shù)據(jù)插入數(shù)據(jù)庫(kù)。
• Use Postman. Postman is a GUI application for working with APIs. Postman allows you to enter values into a GUI and interact with APIs in a drag/drop/click format. once you’ve successfully pulled your data in Postman, you can export the API call in Python and then place that code snippet into your script.
  使用郵遞員。 Postman是用于API的GUI應(yīng)用程序。 Postman允許您將值輸入GUI并以拖放/單擊格式與API交互。 在Postman中成功提取數(shù)據(jù)后，您可以在Python中導(dǎo)出API調(diào)用，然后將該代碼段放入腳本中。
• If you don’t want to deal with any coding, and you have spare budget to work with, I recommend using either Stitch or Xplenty. These are ETL platforms that have built-in integrations with most major APIs, allowing you to move data into your database without any code. The downside to these platforms is the cost — but if you have the budget I highly recommend these as an alternative to maintaining ETL pipelines.
  如果您不想處理任何編碼，并且有多余的預(yù)算可以使用，建議您使用Stitch或Xplenty。 這些ETL平臺(tái)已與大多數(shù)主要API進(jìn)行了內(nèi)置集成，使您無(wú)需任何代碼即可將數(shù)據(jù)移入數(shù)據(jù)庫(kù)。 這些平臺(tái)的缺點(diǎn)是成本–但是，如果您有預(yù)算，我強(qiáng)烈建議將其作為維護(hù)ETL管道的替代方法。

2.了解Web分析，點(diǎn)擊流數(shù)據(jù)和標(biāo)記 (2. Understanding of Web Analytics, Clickstream Data, and Tagging)

It’s more than likely that your website is the primary place you send traffic and clicks with your marketing campaigns. Subsequently, to tell a complete marketing story you need to understand web analytics, clickstream data, and how that data ties to the ads you are creating.

您的網(wǎng)站很有可能是您通過(guò)營(yíng)銷(xiāo)活動(dòng)發(fā)送流量和點(diǎn)擊的主要場(chǎng)所。 隨后，要講述完整的營(yíng)銷(xiāo)故事，您需要了解網(wǎng)絡(luò)分析，點(diǎn)擊流數(shù)據(jù)以及該數(shù)據(jù)如何與您正在創(chuàng)建的廣告聯(lián)系。

Here’s a simple example of some raw click-stream data:

這是一些原始點(diǎn)擊流數(shù)據(jù)的簡(jiǎn)單示例：

You can see there is a timestamp, a unique identifier, the URL, and an SDID column that contains the unique campaign identifier from the URL tracking parameters (The tracking parameters are the values after the ‘?’ in the URL).

您可以看到時(shí)間戳，唯一標(biāo)識(shí)符，URL和SDID列，其中包含來(lái)自URL跟蹤參數(shù)的唯一廣告系列標(biāo)識(shí)符(跟蹤參數(shù)是URL中“？”之后的值)。

When campaigns are created in an ad platform (Facebook, Linkedin, Twitter, etc.) — information from those campaigns or ads needs to tie back to the URL where traffic is being directed. This typically happens with URL tracking parameters, but you’d be surprised how many marketing teams are either a) not tagging anything, or b) tagging ads inconsistently or incorrectly. Any discrepancy between tagging and analytics is going to mean huge gaps in the data that you are trying to collect, which will make it extremely challenging, or impossible to effectively measure marketing campaigns.

在廣告平臺(tái)(Facebook，Linkedin，Twitter等)中創(chuàng)建廣告系列時(shí)-來(lái)自這些廣告系列或廣告的信息需要綁定到定向流量的URL。 URL跟蹤參數(shù)通常會(huì)發(fā)生這種情況，但是您會(huì)驚訝于有多少營(yíng)銷(xiāo)團(tuán)隊(duì)要么a)沒(méi)有標(biāo)記任何內(nèi)容，要么b)不一致或錯(cuò)誤地標(biāo)記了廣告。 標(biāo)記和分析之間的任何差異都將意味著您要收集的數(shù)據(jù)之間存在巨大差距，這將使其變得極具挑戰(zhàn)性，甚至無(wú)法有效地衡量營(yíng)銷(xiāo)活動(dòng)。

Due to the complexity of this process, I always recommend that the analyst/data scientist/whoever is going to be in charge of measurement of marketing programs, be involved with or create the tracking and tagging strategy herself. Typically the marketing team or agency will create the tagging strategy, which often leaves valuable data off the table due to a misunderstanding of the data gathering and cleaning process (for example, using the same unique identifier on multiple advertisements will make it impossible to tell which ad was responsible for the traffic/conversions/etc.). When the analyst herself owns this process, getting clean data at the end will be infinitely easier.

由于此過(guò)程的復(fù)雜性，我始終建議分析師/數(shù)據(jù)科學(xué)家/負(fù)責(zé)市場(chǎng)營(yíng)銷(xiāo)計(jì)劃的人員，自己參與或創(chuàng)建跟蹤和標(biāo)記策略的人員。 通常，營(yíng)銷(xiāo)團(tuán)隊(duì)或代理商將創(chuàng)建標(biāo)記策略，由于誤解了數(shù)據(jù)收集和清理過(guò)程，該策略經(jīng)常會(huì)將有價(jià)值的數(shù)據(jù)從表格中刪除(例如，在多個(gè)廣告上使用相同的唯一標(biāo)識(shí)符將使您無(wú)法分辨出哪個(gè)標(biāo)記廣告負(fù)責(zé)流量/轉(zhuǎn)化/等)。 當(dāng)分析人員自己擁有這個(gè)過(guò)程時(shí)，最后獲得干凈的數(shù)據(jù)將變得非常容易。

Image by author圖片作者

The above process is something that rarely happens, even with marketing teams at the biggest most technical companies. Typically, the marketers will use Adobe Analytics or Google Analytics dashboards OR they’ll default to platform-specific analytics. This can work sometimes — especially if you’re primarily an eCommerce company. Problems arise when you want all the data tied together — which currently can only be done by an analytics professional who understands the dynamics of what I’ve discussed above. B2B businesses are especially susceptible to this problem as many B2B sales take place in conference rooms and not online. So to actually measure marketing impact on sales you have to somehow connect marketing data to CRM data — which is incredibly challenging without the right analytics talent — i.e. someone whos learned these critical skills.

即使有最大的技術(shù)公司的營(yíng)銷(xiāo)團(tuán)隊(duì)，上述過(guò)程也是很少發(fā)生的。 通常，營(yíng)銷(xiāo)人員將使用Adobe Analytics或Google Analytics(分析)儀表板，或者默認(rèn)使用特定于平臺(tái)的分析。 有時(shí)這可以工作-特別是如果您主要是一家電子商務(wù)公司。 當(dāng)您希望將所有數(shù)據(jù)捆綁在一起時(shí)，就會(huì)出現(xiàn)問(wèn)題-目前這只能由了解我上面討論的動(dòng)態(tài)的分析專(zhuān)家來(lái)完成。 B2B企業(yè)特別容易受到此問(wèn)題的影響，因?yàn)樵S多B2B銷(xiāo)售都是在會(huì)議室而不是在線進(jìn)行的。 因此，要真正衡量營(yíng)銷(xiāo)對(duì)銷(xiāo)售的影響，您就必須以某種方式將營(yíng)銷(xiāo)數(shù)據(jù)與CRM數(shù)據(jù)聯(lián)系起來(lái)-如果沒(méi)有合適的分析人才，這將是非常艱巨的挑戰(zhàn)-即，已經(jīng)學(xué)會(huì)了這些關(guān)鍵技能的人。

3.強(qiáng)大的業(yè)務(wù)和營(yíng)銷(xiāo)領(lǐng)域知識(shí) (3. Strong Business and Marketing Domain Knowledge)

Once you have clean data compiled and integrated, then you’re ready to make that data digestible for marketing and sales teams to consume. This may be the most difficult and nuanced skill of all, as it requires you to move outside of doing technical analytics and data work, and spend time with business and marketing practitioners. You need to have a few things clear before you can create a dashboard or report that can effectively communicate marketing performance:

整理并集成了干凈的數(shù)據(jù)后，就可以準(zhǔn)備好使這些數(shù)據(jù)易于消化，以供營(yíng)銷(xiāo)和銷(xiāo)售團(tuán)隊(duì)使用。 這可能是所有技能中最困難和最細(xì)微的技能，因?yàn)樗竽D(zhuǎn)而從事技術(shù)分析和數(shù)據(jù)工作，并花時(shí)間與業(yè)務(wù)和營(yíng)銷(xiāo)從業(yè)人員聯(lián)系。 在創(chuàng)建可以有效傳達(dá)營(yíng)銷(xiāo)績(jī)效的儀表板或報(bào)告之前，您需要明確一些事項(xiàng)：

• What does the marketing team actually care about? Never assume what your marketing team's priorities are. You need to be very clear with the marketers you work with on what their success metrics are. If they don’t know, then you’ll have to create them and that means knowing the marketing programs and business well enough to convince the teams of KPIs that make sense.
  營(yíng)銷(xiāo)團(tuán)隊(duì)實(shí)際上在乎什么？ 永遠(yuǎn)不要假設(shè)您的營(yíng)銷(xiāo)團(tuán)隊(duì)的工作重點(diǎn)是什么。 您需要與合作的營(yíng)銷(xiāo)人員非常清楚其成功指標(biāo)是什么。 如果他們不知道，那么您就必須創(chuàng)建它們，這意味著要充分了解營(yíng)銷(xiāo)計(jì)劃和業(yè)務(wù)，才能說(shuō)服有意義的KPI團(tuán)隊(duì)。
• What do they consider successful? Even if your marketing team has given you success metrics that they want to look at, they’ll always want to compare it to something — whether that’s an industry-standard performance metric or time driven (i.e. this time last year, last year year-to-date, etc.) Never create a dashboard that doesn’t allow comparison against a standard or time-series.
  他們認(rèn)為成功了什么？ 即使您的營(yíng)銷(xiāo)團(tuán)隊(duì)為您提供了他們想看的成功指標(biāo)，他們也總是想將它與其他指標(biāo)進(jìn)行比較-無(wú)論是行業(yè)標(biāo)準(zhǔn)的績(jī)效指標(biāo)還是受時(shí)間驅(qū)動(dòng)的指標(biāo)(例如，去年的這個(gè)時(shí)候，去年的一年，等等)，切勿創(chuàng)建不允許與標(biāo)準(zhǔn)或時(shí)間序列進(jìn)行比較的儀表板。
• Is failure ok? This is unfortunate, but several marketing teams that I’ve worked with only cared about data if it made their programs look good. You need to know this from the start — is it the intention of the team to measure and improve or just present information? Many teams just want to show stakeholders they spent the budget and drove a bunch of traffic — they don’t really care about making better ads. You’ll be making two entirely different dashboards/reports depending on which type of marketing team you’re working with.
  失敗可以嗎？ 不幸的是，但是我曾與之合作的幾個(gè)營(yíng)銷(xiāo)團(tuán)隊(duì)只關(guān)心數(shù)據(jù)，如果這些數(shù)據(jù)使他們的程序看起來(lái)不錯(cuò)。 您需要從一開(kāi)始就知道這一點(diǎn)-團(tuán)隊(duì)的目的是衡量和改進(jìn)或只是提供信息？ 許多團(tuán)隊(duì)只是想向利益相關(guān)者展示他們花費(fèi)了預(yù)算并吸引了大量流量-他們并不真正在意制作更好的廣告。 根據(jù)要使用的營(yíng)銷(xiāo)團(tuán)隊(duì)類(lèi)型，您將制作兩個(gè)完全不同的儀表板/報(bào)告。

營(yíng)銷(xiāo)人員最好的朋友 (The Marketers Best Friend)

The three skills I’ve outlined above are absolutely critical for doing good marketing analytics. Without them, you’ll be unable to effectively measure and report on your companies marketing programs. With them, you’ll become a rare asset that will quickly make you the marketer's best friend.

我上面概述的三項(xiàng)技能對(duì)于做好營(yíng)銷(xiāo)分析至關(guān)重要。 沒(méi)有他們，您將無(wú)法有效地衡量和報(bào)告公司的營(yíng)銷(xiāo)計(jì)劃。 有了它們，您將成為稀有資產(chǎn)，將Swift使您成為營(yíng)銷(xiāo)人員的最佳朋友。

Questions or comments? You can email me at cwarren@stitcher.tech. Or, follow me on Linkedin at https://www.linkedin.com/in/cameronwarren/

有疑問(wèn)或意見(jiàn)嗎？ 您可以通過(guò)cwarren@stitcher.tech向我發(fā)送電子郵件。 或者，通過(guò)Linkedin在https://www.linkedin.com/in/cameronwarren/上關(guān)注我

I also provide Marketing Analytics services. If you’d like to inquire more, email me directly at cwarren@stitcher.tech or go to http://stitcher.tech/contact/.

我還提供Marketing Analytics服務(wù)。 如果您想了解更多信息，請(qǐng)直接發(fā)送電子郵件至cwarren@stitcher.tech或訪問(wèn)http://stitcher.tech/contact/ 。

If you want help connecting your click-stream and CRM data, check out https://stokedata.com/.

如果您需要幫助來(lái)連接點(diǎn)擊流和CRM數(shù)據(jù)，請(qǐng)?jiān)L問(wèn)https://stokedata.com/ 。

翻譯自: https://towardsdatascience.com/the-3-most-critical-skills-for-marketing-analytics-e68e254908de

營(yíng)銷(xiāo)大數(shù)據(jù)分析 關(guān)鍵技術(shù)

總結(jié)

以上是生活随笔為你收集整理的营销大数据分析 关键技术_营销分析的3个最关键技能的全部?jī)?nèi)容，希望文章能夠幫你解決所遇到的問(wèn)題。

如果覺(jué)得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。